Comprehensive Dual- and Triple-Feature Intersectional Single ...web.stanford.edu/group/dlab/media/papers/fennoNeuron2020.pdfComprehensive Dual- and Triple-Feature Intersectional Single-Vector

NeuroResource

Comprehensive Dual- and
Triple-FeatureIntersectional Single-Vector Delivery of DiverseFunctional Payloads to Cells of Behaving Mammals
Graphical Abstract

Highlights

d Multiple recombinase-dependent expression of 15 new

molecular payloads in single AAVs

d Intersectional Ca2+ imaging, cell labeling, and optogenetic

inhibition or excitation

d Creation and in vivo validation of triple-feature-dependent

viruses (Triplesect)

d Design of a widely adaptable in vivo quantitative expression

tracking system

Fenno et al., 2020, Neuron 107, 1–18September 9, 2020 ª 2020 Elsevier Inc.https://doi.org/10.1016/j.neuron.2020.06.003

Authors

Lief E. Fenno, Charu Ramakrishnan,

Yoon Seok Kim, ...,

Nandini Pichamoorthy,

Alice S.O. Hong, Karl Deisseroth

[email protected]

In Brief

Fenno et al. enable versatile functional

access to cell types defined by the

presence of multiple (2 or 3) features,

creating diverse expression-control logic

contained in single viruses. This result is a

comprehensive toolset enabling multiple-

feature-dependent optogenetic inhibition

and excitation and structure- or activity-

based fluorescence imaging with diverse

new indicators.

ll

mailto:deissero@stanford.�edu

https://doi.org/10.1016/j.neuron.2020.06.003

Please cite this article in press as: Fenno et al., Comprehensive Dual- and Triple-Feature Intersectional Single-Vector Delivery of Diverse FunctionalPayloads to Cells of Behaving Mammals, Neuron (2020), https://doi.org/10.1016/j.neuron.2020.06.003

ll

NeuroResource

Comprehensive Dual- and Triple-FeatureIntersectional Single-Vector Delivery of DiverseFunctional Payloads to Cells of Behaving MammalsLief E. Fenno,1,2,4 Charu Ramakrishnan,2,4 Yoon Seok Kim,2,4 Kathryn E. Evans,2 Maisie Lo,2 Sam Vesuna,2

Masatoshi Inoue,2 Kathy Y.M. Cheung,2 Elle Yuen,2 Nandini Pichamoorthy,2 Alice S.O. Hong,2 and Karl Deisseroth1,2,3,5,*1Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA 94305, USA2Department of Bioengineering, Stanford University, Stanford, CA 94305, USA3Howard Hughes Medical Institute, Stanford University, Stanford, CA 94305, USA4These authors contributed equally5Lead Contact

*Correspondence: [email protected]


SUMMARY

The resolution and dimensionality with which biologists can characterize cell types have expanded dramat-ically in recent years, and intersectional consideration of such features (e.g., multiple gene expression andanatomical parameters) is increasingly understood to be essential. At the same time, genetically targetedtechnology for writing in and reading out activity patterns for cells in living organisms has enabled causalinvestigation in physiology and behavior; however, cell-type-specific delivery of these tools (includingmicro-bial opsins for optogenetics and genetically encoded Ca2+ indicators) has thus far fallen short of versatile tar-geting to cells jointly defined by many individually selected features. Here, we develop a comprehensiveintersectional targeting toolbox including 39 novel vectors for joint-feature-targeted delivery of 13 molecularpayloads (including opsins, indicators, and fluorophores), systematic approaches for development and opti-mization of new intersectional tools, hardware for in vivomonitoring of expression dynamics, and the first ver-satile single-virus tools (Triplesect) that enable targeting of triply defined cell types.

INTRODUCTION

Cellular-resolution investigation of the biology of behaving ani-

mals has leveraged powerful, genetically encoded molecular

tools that utilize visible light for exchange of information with tar-

geted cells, including optogenetic tools to control cellular events

and fluorescent indicators to report cellular signaling and anat-

omy. However, application of these approaches depends on

the ability to selectively express the genetically encoded tools

in well-defined cellular subpopulations—with limitations that

have become increasingly apparent with the rapid progression

of detailed single-cell transcriptomes and connectomes

revealing that cell-type definition with only single features is

inadequate.

A technique termed INTRSECT (intronic recombinase sites

enabling combinatorial targeting) partially addressed this chal-

lenge (Fenno et al., 2014), advancing beyond earlier single-

feature methods for cell-type targeting (which, in the case of viral

tools, was based on enhancer- or anatomy-guided expression of

a single recombinase [e.g., Cre] employed in combination with a

Cre-dependent virus) (reviewed in Fenno et al., 2011; Yizhar

et al., 2011). In contrast, INTRSECT allowed adeno-associated

virus (AAV)-borne molecular payloads to be expressed in cells

based on a doubly specified combination of genetically and/or

anatomically defined parameters by placing two orthogonal re-

combinase (Cre and Flp) recognition sequences in synthetic in-

trons (Fenno et al., 2017). Although only enabled for yellow fluo-

rescent protein (EYFP) and the excitatory channelrhodopsin

fusion protein ChR2-EYFP, the initial proof-of-concept two-

feature INTRSECT has been successfully applied in diverse

experimental settings, including for mapping projection patterns

of certain neuronal subtypes (Chuhma et al., 2018; Poulin et al.,

2018) and for probing the contribution of doubly specified cell

populations to diverse motivated behaviors (Gao et al., 2019;

Lazaridis et al., 2019; Marcinkiewcz et al., 2016; Tovote et al.,

2016). However, the ensuing 6 years have witnessed an explo-

sion in the richness of molecular cell typology, creating a funda-

mental unmet need for a comprehensive toolset beyond ChR2-

EYFP able to investigate the necessity and sufficiency as well

as natural activity of targeted cell populations in multiply defined

neurons, including dual-feature and triple-feature targeting.

Here we report our development of this toolset, driven by an

approach spanning four independent domains of intersectional

expression engineering: (1) design of a pipeline for developing

Neuron 107, 1–18, September 9, 2020 ª 2020 Elsevier Inc. 1

mailto:[email protected]


llNeuroResource


multiple-recombinase-dependent constructs to expand the

repertoire of two-feature targeting to a wide array of commonly

used molecular tools, bringing the total number of validated

intersectional tools from 6 to 45; (2) refinement of the Flp recom-

binase-dependent components of the viral backbone to signifi-

cantly enhance Flp-mediated recombination, widening the range

of intersectional experimental designs available; (3) quantifying

recombinase- and non-recombinase-dependent expression ki-

netics with a novel device for chronic in vivo expression moni-

toring; and (4) the first triple-intersectional viral technology, vali-

dating the resulting tools in vitro and in vivo. These resources

allow detailed and rigorous investigation of the natural and

causal roles of cells defined by the intersection of many geneti-

cally and/or anatomically specified properties.

RESULTS

Development of the Intersectional PipelineINTRSECT combines synthetic introns and dependency on two

recombinases (Cre and Flp) to restrict EYFP or ChR2-EYFP to

cellular populations specified by two features (Figures 1A–1F).

For example, two introns could be inserted into a gene with

two reading frames (initially, an opsin-fluorophore fusion gene).

The starting configuration of the exons and the recombinase

recognition sites (e.g., lox and FRT) determined which logical

combination of Cre and Flp would enable expression (Figures

1D and 1E), and the synthetic introns that contained these re-

combinase sites were reliably removed during mRNA splicing,

allowing functional protein to be expressed (Figure 1F).

To create next-generation intersectional tools with a broad di-

versity of molecular payloads, we began by designing a produc-

tion pipeline (Figures 1G and 1H). We tested this pipeline by

creating novel intersectional viruses for a broad array of

commonly used optical tools, implementing new classes of

payload as well as incorporating vastly more potent capabilities

that have emerged in recent years: fluorescent proteins

(mTagBFP, mCherry, and oScarlet; Figure S1; oScarlet and

sRGECO are variants of mScarlet and jRGECO1a that we engi-

neered to reduce aggregation, as described in Figures S1A–

S1C and S2A–S2E), Ca2+ indicators (GCaMP6f, GCaMP6m,

and sRGECO; Figure S2), excitatory opsins [ChRmine 3.3-

p2a-oScarlet; ChR2(H134R)-mCherry, bReaChES-EYFP, and

ChR2(E123T/T159C)-EYFP; Figure S3], and inhibitory opsins

(iC++-EYFP; NpHR 3.3-p2a-EYFP, and Arch 3.3-p2a-EYFP; Fig-

ure S4). The new pipeline (Figure 1G) was validated through the

generalized success of its informatics-based intron placement in

generating properly spliced products and by the early identifica-

tion of occasional two-intron constructs with spurious splice

products (allowing modification prior to further characterization

in vitro or in vivo; Figure 2).

When necessary, improving splice fidelity required individual-

ized strategies based on the sequence of the mis-spliced prod-

ucts (Figure 2A); for example, we observed a preferred cryptic

splice site in the second exon of bReaChES-EYFP as well as

direct splicing of exon 1 to exon 3 (Figure 2B). In this case, simply

moving the first intron to a 30 secondary candidate splice site was

sufficient to eliminate the cryptic site and exon skipping (Figures

2C and 2D). Separately, mis-splicing at a cryptic site was found

2 Neuron 107, 1–18, September 9, 2020

with NpHR 3.3-p2a-EYFP (Figure 2E); in this case, the sequence

of NpHR did not offer an additional splice site option, and we

were unable to use codon degeneracy to disrupt the cryptic

splice site. We instead leveraged the structure of NpHR

(Kouyama et al., 2010) to consider that the distance (>8 A) of

the residue encoded by the cryptic splice site from the retinal

binding pocket would most likely enable engineering of a site

that would be non-destructive to protein function (Figure 2F,

left); indeed, the introduced mutation W179F did not negatively

influence opsin function (Figure 2F, right; p = 0.9754, unpaired

t test) and successfully resolved mis-splicing (Figure 2G). We

anticipate that our algorithmic approach to addressing mis-

splicing will be a helpful addition to the standard operating pro-

cedure and design manual for INTRSECT implementation

(Figure 2H).

Aside from such rare cryptic splice sites, direct exon 1-to-exon

3 splicing was a frequently observed minor splice variant in two-

intron (three-exon) constructs. We attenuated this phenomenon

with a panel of approaches, including modifying the splice

acceptor polypyrimidine tract C/T content, increasing the intron

sequence length, and increasing the distance between the in-

trons. None of these approaches completely eliminated exon

1-to-exon 3 direct splicing (data not shown); in fact, even original

wild-type (WT) cDNA exhibited the same splice product and

sequence in some cases (including in NpHR 3.3(W179F)-p2a-

EYFP [Figure S4A] and ChR2(H134R)-mCherry [Figure S3G]),

indicative of intrinsic splicing without synthetic introns. These

alternatively spliced variants were minor products that did not

adversely affect functional expression relative to the WT.

Flow cytometry in HEK293 cells revealed operation as ex-

pected for all of the novel Cre AND Flp (Con/Fon) vectors (sche-

matized in Figures 1B and 1E): no expression in the absence of

recombinases, an expression level comparable with the WT

when paired with activating recombinases, and no off-target

expression. The Flp AND NOT Cre (Coff/Fon) constructs ex-

hibited a range of expression associated with tool class: fluoro-

phores and genetically encoded calcium indicators (GECIs) ex-

pressed at moderately reduced levels relative to the WT

(Figures S1E, S1H, S1K, S2G, S2J, and S2M), whereas excit-

atory (Figures S3B, S3E, S3H, and S3K) and inhibitory (Figures

S4B, S4E, and S4H) opsin expression levels were comparable

with those of the WT. Inactivation of Coff/Fon constructs by

co-transfection with Cre as well as Flp was effective at diminish-

ing expression to levels similar to negative control levels. For the

opposite configuration (Con/Foff, Cre AND NOT Flp), co-trans-

fection with Flp as well as Cre diminished expression in all cases;

in some constructs, expression was abolished (i.e., to levels

indistinguishable from those of negative controls), whereas in

others, expression was decreased by more than an order of

magnitude but still detectable 5 days post-transfection. These

constructs thus revealed the potential for transient off-target

expression in cells co-expressing Cre and Flp, which could

occur (for example) if Cre acts before Flp in Con/Foff constructs

because of the known higher efficacy of Cre relative to Flp (Ring-

rose et al., 1998); we address this (Figures S5 and S6) with opti-

mized Flp recombinase-dependent construct design.

Next, regarding function of the resulting correctly assembled

constructs in the pipeline, fluorophores were tested by

single intron intersectional constructs

double intron intersectional constructs

Arch 3.3-p2a-EYFPNpHR 3.3-p2a-EYFP*iC++-EYFP

Inhibitory Opsins

bReaChES-EYFPChRmine 3.3-p2a-oScarlet

ChR2 (H134R)-mCherry

ChR2 (H134R)-EYFP

Excitatory Opsins

ChR2 (ET/TC)-EYFP

*modified forimproved function

*modified forimproved function

Flp-dependentCre-dependent

exon 2

exon 2

exon 1

intron exon 2exon 1

exon 1

Con/Fon

Con/Foff

Coff/Fon

mCherryEYFPBFP

Fluorophores

oScarlet*

exon 3 exon 1

exon 1

exon 2

exon 2

exon 1 exon 3

exon 2

Cre-dependentFlp-dependent

Con/Fon

Con/Foff

Coff/Fon

intron 2intron 1

exon 3

GECIsGCaMP6m

sRGECO*GCaMP6f

exon 2exon 1

exon 2exon 1

active state

Con/Fon initial state

exon 2exon 1exon 2exon 1Coff/Fon initial state Con/Foff initial state

exon 2exon 1exon 2exon 1

Coff/Fon inactivated state Con/Foff inactivated state

+cre & flp

+cre

+cre +flp

+flp

Flp-dependentCre-dependentintron

exonexon

intronexon 2exon 1

exon 2exon 1

initialstate

activestate

mRNA

protein

recombinase activity

intron splicing

translation

exon 3exon 2exon 1

active state

exon 3exon 2exon 1Coff/Fon inactivated state

exon 1exon 2exon 3

Con/Foff inactivated state+cre +flp

exon 1exon 2exon 3

Con/Fon initial state

exon 1exon 2exon 3

Coff/Fon initial stateexon 3exon 2exon 1

Con/Foff initial state+cre & flp

+cre+flp

*

Flp-dependentCre-dependent

intron 1exonexon

intron 2exon

intron 1exon 3exon 1

intron 2exon 2

exon 3exon 1 exon 2

initialstate

activestate

mRNA

protein


intron splicing

translation

A

D

flowcytometryRT-PCR functional

testingdesign &cloning

G

H

B

E

C

F

check for functional defects (in vitro, in vivo)

check expression levelcheck off-target expression

check mis-splicingremove cryptic sitesuse alternative codons

rational engineering

ChRmine 3.3-p2a-oScarlet

WT Con/Fon +Cre +Flp Con/Foff +Cre Coff/Fon +Flp

50 m

V

1 s

1 ms585 nm WT Con/Fon +Cre +Flp Con/Foff +Cre Coff/Fon +Flp

50 m

V

1 s1 s470 nmcurrentinjection

iC++-EYFP

Figure 1. INTRSECT Strategy and Engineering Pipeline

(A and D) INTRSECTmolecular designs for a single open reading frame (ORF; A) and double ORF (D) in three Boolean configurations (Cre AND Flp, Cre AND NOT

Flp, Flp AND NOT Cre). Reagents for each configuration are listed.

(B and E) Activity of Cre and Flp to move the single ORF (B) and double ORF (E) INTRSECT starting configurations (top) to the active (dotted box, center) and

inactivated (bottom) states.

(C and F) From top to bottom: how the initial DNA configurations for single-ORF (C) and double-ORF (F) constructs transition to the active state after recombinase-

dependent rearrangement, mRNA processing that removes introns containing recombinase recognition sites, and translation without addition of an extraneous

sequence (crystal structure in C: GCaMP6m, PDB: 3WLD [Ding et al., 2014]; F: iC++ PDB: 6CSN [Kato et al., 2018]).

(G) Standardized engineering pipeline for production of novel INTRSECT constructs consisting of (left to right) design of intron placement and cloning, RT-PCR to

ensure proper splicing, flow cytometry to assay proper expression, and functional testing (in cultured neurons or HEK cells) to compare with the parent construct.

(H) Electrophysiology in cultured neurons expressing WT, Con/Fon, Con/Foff, or Coff/Fon variants of ChRmine 3.3-p2a-oScarlet (left) or iC++-EYFP (right) with

recombinases.

See also Figures S1–S4.

llNeuroResource


expression in cultured cells in addition to flow cytometry analysis

(Figures S1F, S1I, and S1L). GECIs were transfected into

neuronal primary cultures and functionally assayed by electrical

field stimulation with single-cell resolution fluorescence [Ca2+]iimaging. All GECI constructs (sRGECO, GCaMP6f, and

GCaMP6m) generated reliable signals. Basal expression from

these intersectional constructs was significantly decreased rela-

tive to the parental tool in some cases (Figures S2H and S2K; all

versus WT; sRGECO Con/Fon p = 0.0034, Con/Foff and Coff/

Fon p < 0.0001; GCaMP6m Coff/Fon p = 0.0414, Kruskal-Wallis

test with Dunn’s test); [Ca2+]i signals (dF/F; see Methods) were

greater in cases of lower basal fluorescence (Figures S2H and

Neuron 107, 1–18, September 9, 2020 3

INTRSECT SOP(optogenetics.org)

Design manual(CSHL)

virus production

1. labeling errors2. cross-contamination

3. production failure

application

1. labeling errors2. virus mishandling

3. transgenic limitations4. syringe cross-contamination

amplification


3. equipment failure

A

H

rational protein engineering

0.5

0Max

pho

tocu

rrent

(nA)

5 5

n.s.

WT

W179F

W179FWT TTCTGGTAC

TTCTTTTACF W/F Y

Cryptic splice donor

RetinalRetinal

W179W179

extra-cellular

intra-cellular ion-conducting

pathwayion-conducting

pathway

storage


NpHR-EYFP

* ** *

600 bp

1.5 kb

* *

## #

250 bp 914 bp R

F

intronintronNpHRC EYFPN EYFPCNpHRN

582 bpprimers

NpHR(W179F)-EYFP

primers

500 bp

2 kb

* *

250 bp 914 bp R

F

intronintronNpHRC EYFPN EYFPCNpHRN

582 bpW179F

intron site replacement

219bp 921bp

intronintronbRChC EYFPN EYFPCbRChN

561bp

459bp 681bp


561bpalternativeplacement

initialplacement

240bp

intronreplacement

C

F

600 bp

1.1 kb

* *

## #

219 bp 921 bp R

F


561 bpprimers

bReaChES-EYFPB

E

459 bp 681 bp R

F


561 bpprimers

optimized bReaChES-EYFPD

G

1.5 kb

wtcDNADNA

Con/Fon Con/Foff Coff/FoncDNADNA

wt Con/Fon Con/Foff Coff/Fon

cDNADNA

wt Con/Fon Con/Foff Coff/FoncDNADNA

wt Con/Fon Con/Foff Coff/Fon

mis-splicing?RT-PCR

N

N

Y

Y

Nflow cytometry

alternative site?move

intron site

degenerate codon?codon replacement

rational protein engineering

Y

TAT TAC

?T

A C

Gcorrect mis-spliced

Figure 2. Standardized INTRSECT Design and Implementation

(A) RT-PCR testing and mis-splicing resolution approach for new INTRSECT constructs.

(B and E) Mis-spliced RT-PCR results for INTRSECT bReaChES-EYFP and NpHR3.3-p2a-EYFP. bReaChES-EYFP (B) and NpHR 3.3-p2a-EYFP (E) had major

and minor splice variants from cryptic splicing (denoted by #) and exon 1-to-exon 3 direct splicing (denoted by *).

(C and F) The bReaChES-EYFP intron was moved to an alternative splice site (C). NpHR 3.3-p2a-EYFP did not present options for use of a separate splice site or

degenerate codons. Guided by the published crystal structure, we disrupted the cryptic splice site (F, arrow) by introducing themutationW179F (F, center), which

did not affect opsin function (F, right; p = 0.9754, unpaired t test).

(D andG) These iterations of bReaChES-EYFP (D) and NpHR 3.3-p2a-EYFP (G) generated single spliced products (D) or the correct major product and an exon 1-

exon 3 minor splice variant (G).

(H) To catch errors early during scaling and implementation, we described a protocol for making new INTRSECTs (Fenno et al., 2017) and maintain a standard

operating procedure (http://www.optogenetics.org/intrsect_sop.pdf).

See also Figures S1–S4.

llNeuroResource


S2N; sRGECO Con/Fon p = 0.0161, Con/Foff and Coff/Fon p <

0.0001, GCaMP6f Con/Fon p = 0.0468, Coff/Fon p < 0.0001,

Kruskal-Wallis test with Dunn’s test). Whole-cell electrophysi-

ology of opsin variants under photostimulation was generally

indistinguishable compared with WT/parental constructs (Fig-

ures S3C, S3F, S3I, S3L, S4C, and S4I), with the exception of

Arch 3.3-p2a-EYFP, where Con/Fon and Con/Foff exhibited

moderately diminished photocurrents in culture (p < 0.05 for


both, ANOVA with Dunnett’s test). To explore this effect, we pro-

duced AAVs of the parental and Arch 3.3-p2a-EYFP constructs

and co-injected these with AAVs encoding activating recombi-

nases (Ef1a-Cre, Ef1a-Flp, and Ef1a-Flp-2a-Cre) into the mouse

hippocampus for further evaluation by slice electrophysiology. In

contrast to photocurrents observed after 1 week of cultured

neuron expression, we found that photocurrents with intersec-

tional Arch 3.3-p2a-EYFP tools in slice 4 weeks post-injection

http://www.optogenetics.org/intrsect_sop.pdf

llNeuroResource


(a more natural application setting than cultured neurons) were

indistinguishable from WT Arch 3.3-p2a-EYFP (Figure S4F;

Con/Fon p = 0.3966, Con/Foff p = 0.9286, Coff/Fon p <

0.0001, ANOVA with Dunnett’s test).

This newmolecular engineering pipeline, which identifies sub-

optimal function early in the production process (Figure 2A) and

enables efficient creation of intersectional constructs in silico

that function well out of the gate after cloning (Figure 1G), al-

lowed systematic generation and validation of a broad repertoire

of intersectional viral tools for precision cell-type investigation in

biology and brings the total number to 45; at the same time, this

algorithmic approach was crucial for creating the triple-recombi-

nase-dependent virus described below.

Improvement of the FRT CassetteAlthough flow cytometry data largely matched functional

expression when constructs were paired with correct recombi-

nases, we consistently observed a minor population of cells

with residual expression 5 days after co-transfection of Con/

Foff constructs with Cre and Flp. We hypothesized that this

pattern might result from inefficiency of Flp relative to Cre,

characterized to be an order of magnitude less efficient at

equimolar concentrations in vitro (Ringrose et al., 1998). We

therefore sought to reduce this off-target expression by

screening Con/Foff variants containing modifications of the

Flp-dependent elements for increased sensitivity to Flp-medi-

ated recombination.

The Flp-dependent cassette (Figure S5A, top) utilizes two in-

dependent Flp recognition elements in the double-floxed in-

verted open reading-frame (DIO; Zhang, 2008) configuration

for recombinase-dependent inversion of exons (Zhang, 2008;

Atasoy et al., 2008; Sohal et al., 2009). The original INTRSECT

design utilized the F3 and F5 sequences (Schlake and Bode,

1994), chosen to avoid potential intermolecular recombination

between virus and the genome of transgenic Cre-expressing an-

imal lines, some of which may contain a residual FRT sequence.

Here we used a rational screening approach that started with a

wide range of Con/Foff-EYFP variants (Figure S5A, bottom)

and further modified promising ones. Candidates were screened

by flow cytometry in vitro to evaluate mean EYFP intensity of the

residual population as well as the percentage of the parent pop-

ulation these residuals represent (Figure S5B). We found that re-

placing the F3 site with FRT (FRT/F5) or a modified form contain-

ing an additional 14-bp palindromic sequence (14bp-FRT/F5)

significantly decreased the residual expression signal as well

as the percentage of cells that continued to aberrantly express

EYFP 5 days post-transfection, while maintaining high expres-

sion in the active configuration (co-transfected with Cre alone;

Figure S5C).

We next assayed whether this improvement in function at an

equimolar Flp:Cre ratio was maintained across other Flp:Cre ra-

tios by comparing our original Con/Foff-EYFP with these two

variants and systematically varying the relative amounts of Cre

and Flp. Both candidate plasmids maintained their improved

expression pattern across a wide range of recombinase ratios

(Figures S5D, top and S5E, top). Consistent with our hypothesis

that residual expression is driven by inefficiency of Flp relative to

Cre, ratios of Flp:Cre beyond 1:1 reduced residual expression,

whereas ratios greater than 10:1 contributed marginal improve-

ment as expression neared fitted floor values for mean expres-

sion and fraction of the population with residual expression (R2

mean expression original configuration v1 = 0.8028, variant g =

0.7114, variant o = 0.6921; R2 fraction with residual expression

original configuration v1 = 0.2793, variant g = 0.5848, variant

o = 0.3983). The magnitude of the improvement was equivalent

between these two variants (Figures S5D, bottom and S5E, bot-

tom; all p > 0.25, ANOVA with Sidak’s test). Last, we evaluated

the performance of these two variants in vivo in cohorts co-in-

jected with AAV-Con/Foff-EYFP incorporating the functional

components of variants v1, g, or o and AAV-Cre or AAV-Flp-

2a-Cre. In vivo performance closely parallel our in vitro observa-

tions, with variants g and o outperforming variant v1 and negli-

gible difference between them (Figure S5F; residual signal rela-

tive to v1 p = 0.0009 variant g, p = 0.0008 variant o, ANOVA

with Dunnett’s test). Because these two variants appeared to

be equivalent in vitro and in vivo, we chose to continue with

the FRT/F5-based variant to simplify cloning and decrease

payload size.

We integrated this improved Flp cassette into all Con/Foff

constructs from the comprehensive new intersectional toolbox

and compared the original (F3/F5) and improved (FRT/F5) ver-

sions (Figures S5G–S5I) across all tools. As expected, when

transfected with Cre alone, we found no significant difference

between original and improved versions in the mean signal or

fraction of positive cells (Figure S5G; Con/Foff-EYFP versus

Con/Foff-(FRT/F5)-EYFP, mean signal p = 0.46, fraction of pos-

itive cells p = 0.47, paired t tests). In contrast, when co-trans-

fected with equimolar amounts of Cre and Flp, the FRT/F5

constructs performed significantly better than their F3/F5 coun-

terparts (Figure S5H; Con/Foff-EYFP versus Con/Foff-(FRT/F5)-

EYFP, mean signal p = 0.0069, fraction of positive cells p =

0.010, paired t tests). To assay whether the observed differ-

ences were due to changes specific to Flp activity or to overall

construct expression, we additionally analyzed the data by

normalizing inactive (Cre and Flp) to paired active (Cre alone)

data and found that the consistent improvement seen with

FRT/F5 is largely driven by a Flp-dependent reduction in the re-

sidual fraction of positive cells (Figure S5I; mean signal p =

0.080, fraction of positive cells p = 0.0010). To further assess

the function of Con/Foff-(FRT/F5)-EYFP in vivo, we injected

AAV-Con/Foff-(FRT/F5)-EYFP into the mPFC and dorsal

hippocampus of Ssttm(Cre) (SST-Cre) animals, either alone or

with AAV-Flp (Figures S6A–S6D); 4 weeks after surgery, we

observed robust expression of EYFP when injected alone and

extinguished expression when co-injected with Flp that was

indistinguishable from uninjected WT controls.

We recommend that Con/Foff experimental designs utilize ra-

tios of Flp:Cre that favor Flp, ideally with a 10:1 ratio, which may

be achieved through viral delivery or utilizing Flp transgenic lines

with high expression. Our characterization of the Flp cassettes,

and the function of the intersectional Con/Foff backbone in

particular, illustrates the importance of quantifying and address-

ing potential off-target expression and provides a practical eval-

uation framework to enable wider adoption of the intersectional

expression platform specifically and Flp-dependent constructs

more generally.


A

bp 535/22bp 535/22

bandpass497/16dichroic

525 LP

spectrometer

505 nmLED

(to animal)

filter box

fiberfiber

fiber

fiber

fiber

fiber

I

0

2

1

3

4

5

noneco

ntra

ipsi

+Flpco

ntra

ipsi

+Flp-2a-Cre

contr

aips

i

EYFPco

ntra

ipsi

Con/Fon-EYFP

+Cre e10co

ntra

ipsi

+Cre e12co

ntra

ipsi

fluor

esce

nce

(a.u

. x10

e7)

n.s.

noneco

ntra

ipsi

+Flpco

ntra

ipsi

+Flp-2a-Cre

contr

aips

i

EYFPco

ntra

ipsi

+Cre e10co

ntra

ipsi

+Cre e12co

ntra

ipsi

0

2

1

3

4

fluor

esce

nce

(a.u

. x10

e7)

Coff/Fon-EYFP

***

Con/Foff 2.0-EYFP

0

2

1

3

4

fluor

esce

nce

(a.u

. x10

e7) n.s.

noneco

ntra

ipsi

+Flpco

ntra

ipsi

+Flp-2a-Cre

contr

aips

i

EYFPco

ntra

ipsi

+Cre e10co

ntra

ipsi

+Cre e12co

ntra

ipsi

F

norm

aliz

ed lo

g ex

pres

sion

expression time (d)5 10 15 20 25 30 35 40

1

0.75

0.5

0.25

0e12 INTRSECT

curve fit95% confidence

D

0.10.1 10 100

1

10

100

AUC

(abs

orba

nce)

1integration time (s)

R2 0.9999

Full spectrum(all samples)

528 - 540 nm(within dynamicrange)

1

10

100

AUC

(photons x10e8)

G

PL

Cg

AP +1.8mm

M2

fibertrack

500 µm

EYFP

H

expression time (d)403530252015105

1

0.75

0.5

0.25

0

Con/Fon-EYFP + Flp-2a-Cre Coff/Fon-EYFP + Flp


1

0.75

0.5

0.25

0

Titere12e12 + e12e11 + e10e12 + e12

t50% expression days (95% CI; range)

4.23 (4.05 - 4.43; 3.32 - 5.12)4.98 (4.66 - 5.35; 3.59 - 7.05)4.59 (4.45 - 4.73; 4.15 - 5.19)5.79 (5.32 - 6.36; 3.79 - 10.26)

R2

0.974770.931220.986340.84837

n6666

b0.163810.139200.151170.11967

n.s.

Virus

Con/Fon -EYFP + Flp-2a-CreCon/Foff -EYFP + CreCoff/Fon -EYFP + Flp

EYFP

Con/Foff-EYFP + Cre


1

0.75

0.5

0.25

0

norm

aliz

ed lo

g ex

p res

sion


1

0.75

0.5

0.25

0

EYFP

e11 INTRSECT + e12 Cree11 INTRSECT + e11 Cree11 INTRSECT + e10 Cre


18.29 (17.51 - 19.14; 14.34 - 22.14)21.52 (20.14 - 23.11; 15.5 - 30.46)19.82 (19.21 - 20.46; 17.95 - 22.45)25.03 (22.99 - 27.47; 16.39 - 44.35)

t95% expression days (95% CI; range)

e12 INTRSECT


e12 INTRSECT


e12 INTRSECT


B C

0

0.1

0.2

0.3

0.4

0.5

0.6

abso

rban

ce (a

.u.)

0.4

0.8

1.2

1.6

1.8

photons (x10e8)400 450 500 550 600 650 700

wavelength (nm)

1000 ms900 ms800 ms700 ms600 ms500 ms400 ms300 ms200 ms100 ms

fluorescein slide

0 500 100002 1468

AUC

(abs

orba

nce)

AUC

(photons x10e9)

integration time (ms)

R2 0.99992

400 450 500 550 600 650 700

0.2

0.4

0.6

0.8

1

abso

rban

ce (a

.u.)

0

0.5

1

1.5

2

2.5

3

3.5

photons (x10e8)

wavelength (nm)

20 s10 s

5 s3 s1 s

0.5 s0.1 s

0.05 s

Con/Fon-EYFP + Flp-p2a-Cre day 5528 nm - 540 nm

E

5 10 15 20 25 30 35 40

10

100

1

10000

1000

expr

essi

on s

core

expression time (d)

day 5

Figure 3. Chronic Monitoring of Viral Expression: Equivalent INTRSECT and WT Expression Kinetics

(A) Expression monitoring device: a light emitting diode (LED) light source fed into a filter cube and coupled to a visible-wavelength spectrometer for

emission detection.

(legend continued on next page)

llNeuroResource



llNeuroResource


Modeling Intersectional Virus Kinetics In Vivo via aNovelSpectroscopy DeviceWe next turned our attention to characterizing the in vivo dy-

namics of intersectional viruses to ask whether recombinase

dependence would change rate of expression in vivo and to pro-

vide guidance regarding experimental design incorporating

intersectional viruses. Expression-kinetics parameters of non-

recombinase-dependent AAV8 have been characterized previ-

ously by histology (Klein et al., 2006; Reimsnider et al., 2007),

showing that expression velocity peaks at 2–3 weeks, followed

by a plateau. To our knowledge, there is no published study

that thoroughly quantifies the expression time course of

commonly employed and/or new optical tools in vivo; this knowl-

edge void could result in heterogeneity among experimental de-

signs employing AAV-delivered molecular tools. Because

behavioral experiments are frequently conducted over days to

weeks, experimental designs that do not wait for peak viral

expression may result in recruitment of changing populations

(more or fewer) neurons over time.

To address this data void, we designed (for broad applicability

and availability) a robust and inexpensive device for adoption in

laboratories across biology (even without infrastructure for mi-

croscopy or imaging, requiring only a simple visible-wavelength

spectrometer) to track expression dynamics, using off-the-shelf

components to assay fluorescence emission from fluorophore

expression in vivo (Figure 3A). The relationship between the

spectrometer’s tunable integration time and counted photons

was verified to be linear for a fluorescein slide (Figure 3B; R2 =

0.9999) and in vivo virally expressed EYFP (Figures 3C and 3D;

R2 = 0.9999). To assay expression over weeks, we chose to

follow measurements across a wide range of integration times

to enable sensitivity to low expression (with longer integration

times) while maintaining the ability to quantify high expression

(with shorter integration times). The resulting expression score

metric (Figures 3E) spanned many orders of magnitude; to pool

data across animals, we log-transformed and normalized scores

(Figure 3F) to model expression (Figure 3G and 3H).

We applied this approach to the three dual-parameter, inter-

sectional logical configurations to characterize expression ki-

(B) Linear input-output relationship between total counted photons (area under t

time (R2 = 0.9999); spectrometer absorbance (range 0–1) and absolute photons

(C–G) Exemplar data: animal co-injected in the mPFC with AAV-Con/Fon-EYFP

(C) A wide range of spectrometer integration times ensures a continuous dynam

expression through late/strong expression.

(D) Linear relationship between AUC and integration time in dynamic range of th

(E) Expression score: normalizing the AUC to integration time and averaging all exp

the time point from (C) and (D) is noted (arrow).

(F) Viral expression kinetics model: fit to y = 1-e(-bx); y, normalized log expressio

expression curve fit; dashed, 95% confidence of fit; b = 0.12715, R2 = 0.9474).

(G) Chronic viral monitoring does not require components beyond typical optoge

(H) Comparison of WT EYFP expression versus all three INTRSECT logical expres

titers of Cre are initially expressed but cause toxicity over time (Con/Foff-EYFP

chronic monitoring. Expression kinetics between INTRSECT and non-INTRSECT

Con/Fon p = 0.4775, WT and Con/Foff p = 0.7728, WT and Coff/Fon p = 0.1380

(I) Comparison of in vivo expression of all INTRSECT AAV-EYFP variants co-in

fluorescence (6 weeks of expression). There was no difference between expressio

Con/Foff-EYFP (p = 0.2559, unpaired t test). Coff/Fon-EYFP expression was low

unpaired t test).

See also Figures S5 and S6.

netics and compare them with WT EYFP. We prepared cohorts

of animals injected with AAV-EF1a viruses delivering EYFP,

Con/Fon-EYFP + Flp-2a-Cre, Con/Foff-(FRT/F5)-EYFP + Cre,

or Coff/Fon-EYFP + Flp (Figure 3H). Expression of EYFP was

measurable 2 days post-injection and rapidly increased over

2 weeks, reaching 95% of maximal expression between weeks

2 and 3. Intersectional viruses co-injected with recombinases

exhibited similar expression kinetics; expression rate constants

for Con/Fon-EYFP, Coff/Fon-EYFP, and Con/Foff-EYFP did

not differ significantly from non-recombinase-dependent control

EYFP (Figure 3H, column b; all versus WT, WT b = 0.1638 ±

0.01148, Con/Fon b = 0.1392 ± 0.01372 p = 0.4775, Con/Foff

b = 0.1512 ± 0.005248 p = 0.7728, Coff/Fon b = 0.1197 ±

0.01678 p = 0.1380, ANOVA with Dunnett’s test, all ± SEM).

We noted a decrease in the fluorescence of Con/Foff-EYFP

with high titers of Cre recombinase (after 26 days for 1 3

10e11 and after 14 days for 1 3 10e12; Figure 3H), suggesting

that high viral expression of Cre is toxic. This toxicity was not

observed with Cre at a titer of 1 3 10e10 or with Flp or Flp-2a-

Cre at a high titer (1 3 10e12). Separate cohorts with co-injec-

tions of lower titers of Con/Foff-EYFP and Cre at 1 3 10e12

confirmed that this toxicity was a result of Cre expression and

not intersectional virus toxicity (Figure S6E).

We also evaluated animals injected with Con/Foff-EYFP + Flp-

2a-Cre or Coff/Fon-EYFP + Flp-2a-Cre. In the Coff/Fon-EYFP

cohort, we found no transient, off-target expression prior to

Cre inactivation of expression, highlighting the utility of chronic

viral expression monitoring. In the Con/Foff-EYFP cohort, we

found that the time course of low-level off-target expression

with approximately equimolar expression of Cre and Flp peaked

at approximately 2 weeks and slowly decayed to approximately

80% of maximum log expression at 6 weeks (data not shown).

Finally, we used the cohort as an opportunity to confirm expres-

sion profiles of these intersectional viruses using a sensitive re-

porter (EYFP) and efficient actuator (viral recombinase). We in-

jected separate control cohorts with EYFP intersectional

viruses paired with no recombinase, AAV-Cre, AAV-Flp, or

AAV-Flp-2a-Cre. We injected Cre at 1 3 10e10 and 1 3 10e12

titers to confirm our previous observations of toxicity with high

he curve [AUC]) of the band-pass-filtered signal and spectrometer integration

are shown.

and AAV-Flp-2a-Cre.

ic range (color) of non-zero and non-saturated (gray) signals from early/weak

e spectrometer is maintained in vivo (colors as in C; R2 = 0.9999).

ression scores for a given time point within the spectrometer’s dynamic range;

n; x, days; b, rate constant (blue dots, within-animal expression scores; red,

netic experiments (here, a 200-mm fiber).

sion variants of EYFP co-injected with the indicated recombinase viruses. High

+ Cre, green and orange dots), which would not have been apparent without

EYFP viruses are equivalent (comparison of rate constant b between WT and

, n = 6 animals per condition, ANOVA with Dunnett’s test).

jected with all combinations of AAV recombinases as assayed by confocal

n of WT EYFP and Con/Fon-EYFP (p = 0.7615, unpaired t test) or WT EYFP and

er than WT EYFP (EYFP 2.41 3 10e7 a.u. versus 8.96 3 10e6 a.u., p = 0.0003,


xDIO-EYFP

+Flp

+VC

re+D

re+S

Cre

+Cre

fDIO dDIO scDIO vcDIO

side

sca

tter (

a.u.

)

cDIO

+Cre +Dre+Flp +SCre +VCreEYFP negative

50

0

coun

t

50

0

50

0

50

0

50

0

A

EYFP intensity10

110

210

310

510

4

EYFP intensity10

110

210

310

510

4

EYFP intensity10

110

210

310

510

4

EYFP intensity10

110

210

310

510

4

EYFP intensity10

110

210

310

510

4

μm

cDIO-EYFP fDIO-EYFP vcDIO-EYFP+

Cre

+ Fl

p+

VCre

EYFPcDIO-EYFP fDIO-EYFP vcDIO-EYFP

+ C

re+

Flp

+ VC

re

EYFP

M1

M2

Cg

PL

IL

500 μm 50 μm

B

AAV-recombinase+ AAV-DIO-EYFPinto mPFC

+1.8 mm

M2Cg

PL

IL

Bregma +1.8 mm

rAAV-Flp+ AAV-fDIO-EYFP

VTA

Bregma -3.3 mm

AAV-fDIO-EYFP

2 weeksmPFC VTA

rAAV

-Flp

+AA

V-fD

IO-E

YFP

rAAV

-VC

re+

AAV-

vcD

IO-E

YFP

4 weeksmPFC VTA

250 μm 250 μm 50 μm

C

rAAV-recombinase+ AAV-DIO-EYFPinto mPFC

+1.8 mm

AAV-DIO-EYFPinto VTA

-3.3 mm

(legend on next page)

llNeuroResource



llNeuroResource


viral Cre titers. As expected, we saw consistent high expression

in all viruses when paired with correct activating recombinases

(Figure 3I); it is notable that AAV-Coff/Fon-EYFP expression

was lower than that of control AAV-EYFP at equal viral titers

(p = 0.0003, unpaired t test). As expected, we observed no off-

target expression with AAV-Con/Fon-EYFP or AAV-Coff/Fon-

EYFP with any combination of viral recombinases, whereas co-

infection of Con/Foff-(FRT/F5)-EYFP with an approximately 1:1

ratio of Flp:Cre using AAV-Flp-2a-Cre exhibited modest residual

expression, consistent with prior results (Figures S5D and S5E).

Together, these results underscore our recommendation to use

increased ratios of Flp:Cre for Con/Foff viruses while also

revealing a wider tolerance of recombinase ratios for other

expression logic.

We used these histologic data to further validate the in vivo

viral monitoring device by examining the relationship between

post hoc fluorescence and final in vivo spectroscopy expression

score, taken 6weeks post-injection (Figure S6F;R2 = 0.5123, p <

0.0001, n = 30 animals, Pearson correlation). Chronic expression

profiling in this way may be easily and generally applicable for

biological laboratories and here reveals (at these titers) that re-

combinase dependence does not slow viral expression kinetics,

that viral expression builds non-linearly and plateaus between

the second and third weeks of expression, that intersectional vi-

rus expression is maintained for at least 6 weeks (the longest

period examined), and that incorporation of an additional recom-

binase for next-step three-parameter targeting may be unlikely

to slow experimental timelines.

New Regime of Viral Intersectionality: Triple-Recombinase-Based TargetingAs powerful as two-feature intersectional viruses have been,

many meaningfully defined cellular populations might require

targeting by combinations of several genetic and anatomical pa-

rameters (Poulin et al., 2018). For example, known neuronal pop-

ulations defined by (1) a single genetic feature and (2) multiple

axonal projection features (double-projection neurons; Jinno,

2009) are of substantial interest, although assessing the func-

tional significance of these populations (as with the vast majority

of multiple-feature-defined cell types revealed by the ongoing

revolution in single-cell sequencing) remains beyond the reach

of currently available molecular targeting approaches.

Although triple-recombinase dependence has not yet been

achieved in single viruses, our screening of many recombinases

(e.g., VCre, SCre, and Dre) for orthogonality to Flp and Cre by as-

saying expression of single-recombinase-dependent DIO-

ChR2-EYFP constructs in flow cytometry led to identification of

VCre (Suzuki and Nakayama, 2011) as the most promising po-

tential third recombinase (Fenno et al., 2014). To create the

Figure 4. Identifying and Validating a Recombinase Orthogonal to Cre

(A) Flow cytometry of HEK293 cells co-transfected with combinations of recom

structs (xDIO-EYFP, columns). Cre and Dre showed bi-directional cross-activity

VCre showed the expected robust action on its vcDIO-EYFP partner without any

(B) AAV-Cre, AAV-Flp, and AAV-VCre showed expected robust activity when co-i

of expression in the mPFC; low (left) and high (right) magnification. Needle track

(C) rAAV serotypes of Flp and VCre co-injected with respective AAV-xDIO-EYFP

(left), sparse EYFP expression was observed in the mPFC and VTA, and at 4 we

desired new dimension of multi-feature intersectional targeting,

we therefore tested for generality of this result using cytosolic

DIO-EYFP in place of membrane-localized DIO-ChR2-EYFP

constructs (Figure 4A). Consistent with our previous results, we

observed robust activity of Cre, Flp, Dre, and VCre when trans-

fected along with each respective xDIO-EYFP construct (with

less efficient activation exhibited by SCre); however, bi-direc-

tional, off-target activity between Cre and Dre was observed

as before (Fenno et al., 2014). VCre, in contrast, was orthogonal

to all of the tested recombinases (no indication of cross-activity)

and, thus, found to be potentially best-suited for use in parallel

with Cre and Flp. To test crucial in vivo functionality, we next

generated AAVs for these three recombinases and correspond-

ing xDIO-EYFP constructs and then separately injected all per-

mutations of these constructs into mouse medial prefrontal cor-

tex (mPFC; Figure 4B). After 4 weeks of expression, we found

robust expression in subjects co-injected with the proper combi-

nations of recombinase/xDIO-EYFP and no cross-expression in

improper pairings, confirming that this combination of three re-

combinases is suitable for orthogonal use in vivo.

We then sought to create a triple intersectional construct that

would only be active in cells that co-express Cre AND Flp AND

VCre but not in cells with any other combination of recombi-

nases (Figure 5A). To achieve this goal, we first leveraged the

modular design of INTRSECT to create a hybrid version of

our one-intron and two-intron dual-parameter constructs (Fig-

ures 1A and 1D) by inserting two introns into EYFP (Figure 5B).

We utilized the engineering pipeline described here (Figure 1G)

to iterate through many variants with the introns placed at

different locations (Figure 5C) because incorporating multiple

introns into EYFP itself was non-trivial and resulted in a high

degree of mis-splicing. As expected, splicing efficiency was

accurately reflected in EYFP expression levels of HEK293 cells

quadruple-transfected with the three recombinases and Con/

Fon/VCon-EYFP variants (Figure 5D). Through this iterative

approach, we identified a variant with excellent expression

(hereafter called 33-EYFP). We next performed flow cytometry

of HEK293 cells co-transfected with various combinations of

recombinases and the triple-dependent INTRSECT (Triplesect)

construct and confirmed that expression of 33-EYFP was

limited to cells co-expressing all three recombinases (Fig-

ure 5E). We tested the specificity of 33-EYFP expression in vivo

by injecting the mPFC of mice with AAV-33-EYFP and combi-

nations of AAV recombinases, confirming strong specific

expression of this novel Triplesect virus only in cells co-ex-

pressing Cre, Flp, and VCre (Figure 5F).

Having created a proof-of-concept, triple-recombinase-

dependent EYFP and confirmed its specific, strong expression

in vivo, we next asked whether this targeting approach

and Flp

binase expression constructs (rows) and recombinase-dependent EYFP con-

, and some cross-activity was noted when Cre was paired with scDIO-EYFP.

noted in vitro cross-activity.

njected with their partner AAV-xDIO-EYFP without cross-activity after 4 weeks

was used to identify injection sites in samples without expression.

in the mPFC while also injecting AAV-xDIO-EYFP into the VTA. After 2 weeks

eks, high levels of expression in both sites (right).


F

J

L

A

Cre

VCre

Flp

CreANDFlp

ANDVCre

B

R2

F1

exon 1 exon 2 exon 3

Con/Fon/VCon-EYFP

intron 1 intron 2

EYFPN EYFPCEYFPstart(inactive) DNA

intron 1 intron 2EYFPEYFPN EYFPC

+Cre +Flp +VCre(active) DNA

EYFPspliced mRNA


intron splicing

Cre-dependent Flp-dependent VCre-dependent C F1/R2

cDNADNA cDNADNA cDNADNA cDNADNA cDNADNA

wt variant 1 variant 2 variant 3 variant 4

1000 bp

1650 bp

650 bp

EYFP map

wtvariant 4

cDNA sequence

intron 1 splice site

exon 1/exon 2

GGCAAGCTGCCC

GGCAAGCTGCCCGGCAAGCTGCCC

cDNA sequence

intron 2 splice site

exon 2/exon 3

GTCCAGGAGCGC

GTCCAGGAGCGCGTCCAGGAGCGC

Dvariant 1 variant 2 variant 3 variant 4wt

HEK293 culture transfection (variant +Cre +Flp +VCre)

50 µm

E

alone

+VCre

+Flp

+Cre

+Cre +VCre

+Flp +VCre

+Cre +Flp

+Cre +Flp +VCre

EYFP

negative

50

0

Cou

nt

HEK293 (recombinase + variant 4)

EYFP intensity10

1 102

103

105

104

no recombinasemPFC viral injection (AAV-recombinase + AAV-3x-EYFP)

fluor

esce

nce

(a.u

. x 1

07 )

5

2

1

4

3

0

+Cre +Flp +VCre +Cre +Flp +Cre +VCre +Flp +VCre +Cre +Flp +VCre

50 µm

GF1/R2

cDNADNA cDNADNA

wt 3x-G6m

1650 bp

1000 bp

650 bp

H50

0

Cou

nt

GCaMP6m intensity10

1 102

103

105

104

3x-G6m +Cre +Flp +VCrewt G6m

I

K

novel object exploration onset

2% d

F/F

500 ms

TH-Cre +3x-G6m +Flp +VCre

TH-Cre +3x-G6m +Flp

TH-Cre +3x-G6m +VCre

WT +3x-G6m +Flp +VCre 10%

dF/

F

10 sm

ean

bout

pea

k dF

/F (%

)

10

4

2

8

6

0

dF/F

(%)

time (s)

50

25

0

3x-G6m +Cre +Flp +VCre neuron culture

0 10 20 30 40 50

novel object exploration-200 0 200

-1

0

1

2

3

4

5

mea

n bo

ut d

F/F

(%)

shuffle offset (s)flu

ores

cenc

e (a

.u. x

106 )

4

2

6

0

TH-Cre +3x-G6m +Flp +VCre TH-Cre +3x-G6m +FlpAP -3.3 mm

250 µm

TH-Cre +3x-G6m +VCre WT +3x-G6m +Flp +VCre

Figure 5. Engineering, Optimization, Testing, and In Vivo Function of Three-Recombinase-Dependent Triplesect Constructs(A) Potential intersectional populations available with three-recombinase expression. The Cre AND Flp AND VCre intersectional population is denoted by a central

pattern.

(B) Detailed diagram of EYFP divided into three exons with addition of two introns and recombinase recognition sites (top). The activity of Cre AND Flp AND VCre

reorients exons in the sense direction (center). Introns are removed during RNA processing (bottom), ending with an intact mRNA encoding EYFP; we labeled this

three-recombinase-dependent approach 33-EYFP.

(C and D) We generated multiple 33-EYFP construct variants with different intron placement.

(C) Variants 1–3 spliced poorly, whereas variant 4 spliced efficiently, as verified by sequencing (bottom).

(D) Splicing resultsmirrored expression patterns in HEK293 cells co-transfected with 33-EYFP variants and Cre, Flp, and VCre.We therefore used variant 4 going

forward.

(E and F) No expression of 33-EYFP was observed when any of the three recombinases was missing, as assayed by flow cytometry of HEK293 cells (E) or in

animals injected with 33-EYFP and recombinases (F; n = 1–2 animals per condition).

(G–K) 33 engineering approach applied to the calcium sensor GCaMP6m (33-G6m): a similar pattern of proper intron splicing (G) and lack of off-target

expression by flow cytometry of HEK293 cells (H; coloring as in E).

(legend continued on next page)

llNeuroResource



llNeuroResource


(Triplesect) would be generalizable and would, for example,

allow creation of a 33-GCaMP6m. As with 33-EYFP, 33-

GCaMP6m spliced efficiently (Figure 5G) and was only ex-

pressed when co-transfected with all three recombinases (Fig-

ure 5H). We verified functionality of 33-GCaMP6m in cultured

neurons (Figure 5I). To assess in vivo function and specificity of

our 33 expression platform, we next co-injected AAV-33-

GCaMP6m and combinations of recombinases into the ventral

tegmental area (VTA) of THtm(Cre) (TH-Cre) transgenic mice and

inserted 400 mm fiberoptic implants to conduct fiber photometry

during novel object exploration (Gunaydin et al., 2014). Consis-

tent with our previous Triplesect results (Figure 5F), robust

Ca2+ signals were observed in animals co-expressing all three

recombinases (Figure 5J, orange) but not in animals expressing

only two recombinases (Figure 5J, green, purple, and blue). 33-

GCaMP6m animals expressing all three recombinases showed

consistent, time-locked Ca2+ transients beginning immediately

prior to periods of novel object exploration, which was not

seen in animals expressing only two of these recombinases (Fig-

ure 5J, blue bars, Figure 5K). Histology (Figure 5L) of 33-

GCaMP6m animals was consistent with the photometry,

showing no expression in control subjects but the expected

VTA expression pattern in three-recombinase animals. The Tri-

plesect approach and the intron-engineering pipeline described

here thus define a generalizable approach to creation and testing

of versatile tools for probing cell types with a new dimension of

precision and specificity in vivo and in behaving animals.

DISCUSSION

Experimental capabilities for cell-type-specific delivery of func-

tion-testing tools (such as genetically encoded Ca2+ indicators

and microbial opsins) have not kept pace with the insights and

opportunities afforded by single-cell transcriptomics and con-

nectomics. As a result, the ongoing revolution in cellular-reso-

lution molecular phenotyping has remained largely passive

and descriptive. Here, we (1) developed a comprehensive

toolbox for precision viral targeting of cells based on dual-

parameter logic using a standardized algorithmic engineering

pipeline, bringing the total number of INTRSECT viruses to

45, now including a wide array of optically modulated molec-

ular payloads; (2) demonstrated improved recombinase effi-

ciency for specific logical configurations found through a

rational screen utilizing functional criteria; (3) implemented

(I) In vitro function of quadruple-transfected (with Cre, Flp, and VCre) neurons exp

G6m versus WT G6m showed reliable function but with a reduced basal fluoresce

signal-to-noise ratio: 25.01 ± 2.733 versus 15.72 ± 1.66, p = 0.0080; dF/F: 0.2646

48.05 versus 2141 ± 337.2, p < 0.0001; tau: 1.15 ± 0.1112 versus 1.736 ± 0.1179,

(J–L) Co-infection of the TH-Cre mouse VTA with 33-G6m and separate viruses

exploration (J; orange) but not in animals co-infected with 33-G6m and combina

(K) Average Ca2+ signal from traces in (J) time-locked to onset of novel objec

combinases but no signal in two-recombinase controls (total object interactions o

shuffling the bout onset timeswhilemaintaining the bout structure in this same trac

Right: average peak signal from traces in (K) as well as additional three-recombina

recombinase control p < 0.0001, ANOVA with Dunnett’s test).

(L) Exemplary slice images (left) and quantification (right) frommice in (J): expressi

not in two-recombinase control mice (Cre and Flp and VCre active condition vers

sections/mouse, 1 mouse in each of 3 controls, 3 mice in triple-recombinase).

in vivo viral expression-monitoring hardware suitable for wide-

spread use in biology laboratories; and (4) described the first

triple-parameter single-virus targeting logic through creation

of Triplesect.

The Intersectional PipelineBy creating the pipeline, we sought to deliver intersectionally

targetable tools (Fenno et al., 2011, 2014) covering a wide range

of actuation wavelengths, biophysical mechanisms, and tempo-

ral kinetics, including fluorophores ranging from blue to red, mul-

tiple excitatory opsins paired with different fluorophores, a range

of inhibitory opsins, and Ca2+ indicators with different excitation

wavelengths and temporal dynamics. The engineering pipeline

was designed to reveal flaws early (including minor splice vari-

ants and inefficiency of Flp relative to Cre; Figures 2 and S5)

and ensure that each tool would function appropriately. In the

work reported here, splice variants were corrected or, where

persistent, demonstrated to have no adverse effects on function.

The relative inefficiency of Flp constrains the utility of all

combinatorial systems using Flp, motivating our substantial

effort to screen for and improve Flp-dependent components of

the intersectional backbone (complementing previous work to

improve the function of Flp recombinase itself; Buchholz et al.,

1998; Raymond and Soriano, 2007). Although this screen yielded

an approximately 20% improvement in Flp efficiency in the FRT/

F5-based constructs compared with our initial design, it is likely

that more efficient Flp-responsive elements (e.g., FRT sequence

variants) or improved Flp recombinase variants may be devel-

oped in the future to further expand the utility of the intersectional

virus approach. Additional approaches not yet tested include

decreasing efficiency of Cre (possibly through the use of less effi-

cient lox sequence variants to match the recombination kinetics

of the two recombinases; Ringrose et al., 1998) and incorpo-

rating novel recombinase recognition sites that are less prone

to spontaneous recombination (Fischer et al., 2019).

Ongoing Within-Mouse Expression Tracking In Vivo

The known difference in efficiency between recombinases moti-

vated our development of a robust device that could be used

broadly across biological laboratories, including groups not

specialized in optical methodologies. To test whether factors

such as recombinase dependence of intersectional tools or var-

iable recombinase efficiency changed the rate of payload

expression in vivo, we developed a novel, inexpensive fiberoptic

ressing 33-G6m with electrical field stimulation. Population comparison of 33-

nce level (time to peak: 0.3727 ± 0.03938 versus 0.282 ± 0.01359, p = 0.0045;

± 0.03951 versus 0.0466 ± 0.007635, p < 0.0001; basal fluorescence: 280.6 ±

p = 0.0001. 33-G6m, n = 32; WT, n = 43; all mean ± SEM; Mann-Whitney test).

encoding Flp and VCre gave rise to robust calcium signals during novel object

tions of only two recombinases (green, purple, and blue).

t exploration (left) shows a consistent signal from 33-G6m with all three re-

range trace = 32, green trace = 27, purple trace = 32, blue trace = 42). Center:

e showed that the observed signal was not due to random fluctuation (p < 0.01).

se-expressing animals (Cre and Flp and VCre active condition versus each two-

on of 33-G6m only (n = 3mice) when all three recombinases are expressed but

us each two-recombinase control p < 0.05, ANOVA with Dunnett’s test; n = 3


llNeuroResource


device for chronic expression monitoring, using off-the-shelf

commercially available components. More complex larger-fiber

neuroscience-specialized systems (e.g., fiber photometry hard-

ware; Gunaydin et al., 2014) could acquire similar data; however,

the system described here is likely to be more broadly useful

across biological systems spanning organisms and tissues and

will be generally useful for screening experimental subjects prior

to behavioral experiments to assay steady-state expression,

proper interventional (optogenetic) fiber placement, and/or virus

efficacy (especially in unique and valuable individual subjects

such as non-human primates; Diester et al., 2011).

Our finding that recombinase- and non-recombinase-depen-

dent AAV8 expression tracked previous histology-based expres-

sion profiling (Klein et al., 2006; Reimsnider et al., 2007) validated

this approach and provided useful quantitative new tool charac-

terization. Importantly, recombinase dependence of the inter-

sectional tools at these viral titers was found to not slow expres-

sion in vivo. Additional characterization of other serotypes,

payloads, and titers will build further knowledge of critical (and

generally unknown) parameters of viral expression in organs

and species across biology.

TriplesectTo extend targeting resolution to three-parameter dependence,

VCre (Suzuki andNakayama, 2011)waschosen over other tested

recombinases for superior orthogonality toCre andFlp inmultiple

expression platforms, with multiple payloads, and as evaluated

with two separate assays for cross-expression (one in vitro and

the other in vivo); the bi-directional cross-activity of Dre (with

Cre) as well as the poor efficiency of SCre were consistent with

prior work (Weinberg et al., 2017). Here the demonstration of

three-recombinase-dependent viral expression (Triplesect)

marks a new level of viral targeting specificity (in vitro or in vivo);

going forward, additional recombinase/recombination sites,

including Nigri/nox and Panto/pox (Karimova et al., 2016), may

show promise as potential additional orthogonal recombinases

for complementary targeting approaches.

Animal Platforms for Intersectional BiologyTriplesect and the intersectional tool set leverage the expanding

availability of Flp recombinase and other transgenic mouse lines

(Daigle et al., 2018; Karimova et al., 2018; Madisen et al., 2015;

Plummer et al., 2015). From an experimental design perspective,

single-virus targeting offers multiple advantages over use of

multiply crossed, intersectional transgenics. First, viruses are

restricted in time and space, avoiding potential off-target

expression because of transient gene expression during devel-

opment and allowing expression only within specified anatomic

regions of interest (Poulin et al., 2018). In addition, viruses also

afford an opportunity for simultaneous expression of multiple in-

tersectionally expressed, molecular tools in the same anatomic

space for co-labeling, an approach that would generally require

a prohibitive animal husbandry effort to replicate with a purely

transgenic approach. Last, the viral tools provide a vast increase

in resource efficiency because the dissociation of cell targeting

and molecular tool identity allows a single recombinase-ex-

pressing transgenic mouse line to be paired with an unlimited va-

riety of virally delivered tools.


The intersectional viruses described here are designed to be

integrated with any combination of molecular tool and recombi-

nase expression. However, one limitation in fully integrating the

diversity of Flp-expressing animal lines with intersectional tools

is the lack of a centralized database of these animal lines. To

facilitate a wide range of two-parameter (e.g., Cre/Flp double-

transgenic) and three-parameter Triplesect experimental de-

signs (e.g., Cre/Flp double transgenics with additional recombi-

nases delivered by local or retrograde virus; Figure 4C), we

created such a database (Table 1). We searched academic

publications, commercial mouse repositories, and publicly

funded transgenic efforts to inventory currently reported and still

unpublished Flp-expressing mouse lines. We identified 60

mouse lines that represent a total of 47 separate genetic

drivers and one transgenic rat line. Of these, five different

lines have already been used experimentally with INTRSECT

viruses. To ensure the ongoing usefulness of this resource,

we created a web-based version that includes the ability for

researchers to submit data (http://www.optogenetics.org/

flp_lines.html).

General Considerations for Experimental Design withIntersectional VirusesAAV has become the standard for gene transduction in vivo,

including in neuroscience because of the relative safety of the

DNA genome and ease of production and engineering. A feature

of AAVs salient to experimental design is viral tropism—the abil-

ity of viruses to transduce a subset of cells based on the expres-

sion of a viral receptor specific to viral serotype—here referring

to the capsid protein that interacts with the receptor. AAVs are

now routinely produced in a number of serotypes; throughout

this study, we used AAV8 and a synthetic retro-AAV (evolved

to transduce axon terminals; Tervo et al., 2016). The identity of

cells infected by a particular AAV is affected by AAV serotype

(Aschauer et al., 2013; Burger et al., 2004; Davidson et al.,

2000), developmental time point (Chakrabarty et al., 2013), ani-

mal strain (Hordeaux et al., 2018), and viral titer (Nathanson

et al., 2009). Despite the critical role of AAVs in modern biology,

a broad accounting of cell subtype transducibility influenced by

these parameters has not been described, although it is conceiv-

able that this information may be extracted (Cristinelli and Ciuffi,

2018) from existing transcriptome datasets (Ecker et al., 2017)

utilizing a combination of AAV transduction and single-cell

RNA sequencing (RNA-seq).

For viral and transgenic intersectional approaches, imperfec-

tions in reagents (e.g., specificity and sensitivity of recombinase

expression, an important consideration even between separate

knockin animal lines encoding the same transgene at the same

locus; Madisen et al., 2010) are multiplicative, and independent

knowledge that the selected viral tropism and recombinase

expression pattern appropriately recruit the subpopulation of in-

terest is critical. Control experiments co-injecting viruses that

require only Cre or only Flp to be activated and assaying the

expression pattern by histological evaluation is one way to

obtain this critical validation. Characterization of recombinase

expression patterns using only antibody or in situ data on the re-

combinase itself will not give information about viral transducibil-

ity but still serves as an important complementary control.

http://www.optogenetics.org/flp_lines.html


Table 1. Transgenic Flp Mouse Lines for Intersectional Biology

Driver Flp Gene Genetics Source Original Citation INTRSECT Citation

Actb FlpE A Jackson: 003800 Rodrıguez et al., 2000 N/A

Atoh1 FlpO-ER A [email protected] Wojcinski et al., 2019 N/A

Bhlhe22 FlpO C [email protected] Cai et al., 2016 N/A

Cag FlpE A RBRC: 10707, 10708 Kanki et al., 2006 N/A

Cag FlpE-ER A no longer maintained Hunter et al., 2005 N/A

CamK2a FlpE N/A Taconic, not available N/A N/A

Cdx2 FlpO A Jackson: 030288 Abraira et al., 2017 N/A

Cdx2 FlpO A [email protected] Britz et al., 2015 N/A

Crh FlpO D Jackson: 031559 unpublished N/A

DbH FlpO C Jackson: 033952 Robertson et al., 2013 N/A

DbH FlpO C MMRRC: 41577 Sun and Ray, 2016 N/A

DbH FlpO E MMRRC: 41575 Sun and Ray, 2016 N/A

Dlx5/6 Myc-FlpE A Jackson: 010815 Miyoshi et al., 2010 Christenson Wick et al., 2019

EF1A Flp A RBRC: 01251, 01252 Takeuchi et al., 2002 N/A

Fezf2 FlpO E [email protected] Matho et al., 2020 N/A

Gad2 FlpO D [email protected] Alhadeff et al., 2018 N/A

GFAP FlpO A Jackson: 33116 Hara and Verma, 2019 N/A

Gsh2 FlpE D CARD: 2114 S. Esumi, personal

communication

N/A

Hcrt FlpO E [email protected] Chowdhury et al., 2019 N/A

Hoxb8 FlpO A EMMA: 11094 Zhang et al., 2018 N/A

Htr3a FlpO D Jackson: 030755 Schuman et al., 2019 N/A

Lbx1 FlpO C [email protected] Bourane et al., 2015 N/A

mGluR6 FlpO A RBRC: 09715 unpublished N/A

MMTV FlpO A [email protected] L€uond et al., 2019 N/A

Mrgprb4 EGFP-2a-FlpO F Jackson: 021078 Vrontou et al., 2013 N/A

Nestin FlpO-ER A [email protected] Lao et al., 2012 N/A

Nkx2.1 FlpO D Jackson: 28577 He et al., 2016 N/A

Nphs1 FlpO A [email protected] Goldberg et al., 2010 N/A

Npy FlpO D Jackson: 030211 Daigle et al., 2018 N/A

Pdx1 FlpO A [email protected] Schonhuber et al., 2014 N/A

Pdx1 FlpO C [email protected] Wu et al., 2017 N/A

Pet1 FlpE A [email protected].

harvard.edu

Jensen et al., 2008 N/A

Pgk1 FlpO A Jackson: 011065 Wu et al., 2009 N/A

Phox2b FlpO B Jackson: 022407 Hirsch et al., 2013 N/A

Prkcd FlpO E [email protected] B. Li, personal communication N/A

PlexinD1 FlpO E [email protected] Matho et al., 2020 N/A

Plxnd1 dgFlpO D [email protected] Daigle et al., 2018 N/A

Pomc FlpE A no longer maintained Vooijs et al., 1998 N/A

Pvalb FlpE E Jackson: 021191 Madisen et al., 2015 N/A

Pvalb FlpO E Jackson: 022730 Madisen et al., 2015 Hafner et al., 2019

Pvalba FlpO E [email protected] Yu et al., 2018 N/A

Rasgrf2 dgFlpO E Jackson: 029589 Daigle et al., 2018 N/A

Rorb FlpO D Jackson: 029590 Daigle et al., 2018 N/A

Rorb FlpO E [email protected] Daigle et al., 2018 N/A

Slc17a6 FlpO D Jackson: 030212 Daigle et al., 2018 N/A

Slc32a1 FlpO D Jackson: 031331 Daigle et al., 2018 N/A

(Continued on next page)

llNeuroResource






















Table 1. Continued

Driver Flp Gene Genetics Source Original Citation INTRSECT Citation

Slc32a1 FlpO E Jackson: 029591 Daigle et al., 2018 Lazaridis et al., 2019

Slc6a3 FlpO D Jackson: 033673 unpublished N/A

Slc6a4 FlpO D Jackson: 033674 unpublished N/A

SST FlpO D Jackson: 028579 He et al., 2016 Clemente-Perez et al., 2017;

Cummings and Clem, 2020;

Fadok et al., 2017;

Fenno et al., 2014;

Yu et al., 2017

Tbr2 FlpO-ER E [email protected] Matho et al., 2020 N/A

Tbx21 FlpE-ER D Graham.lord@manchester.

ac.uk

Gokmen et al., 2013 N/A

TH FlpO E MMRRC: 50618 Poulin et al., 2018 Chuhma et al., 2018

Poulin et al., 2018

Mingote et al., 2019

TH FlpO A RBRC: 5168-5171 Imayoshi et al., 2012 N/A

Trf FlpE N/A Taconic, not available N/A N/A

Tshz1 FlpO E [email protected] B. Li, personal communication N/A

VGlut2 FlpO D [email protected] Alhadeff et al., 2018 N/A

VIP FlpO D Jackson: 028578 He et al., 2016 N/A

Wnt1 Flp(F70L) A no longer maintained Dymecki and Tomasiewicz,

1998

N/A

Wnt1 FlpE-ER A no longer maintained Hunter et al., 2005 N/A

Wnt1 FlpE A [email protected].

harvard.edu

Awatramani et al., 2003 N/A

Transgenic lines were located via a range of search methods. Lines no longer maintained are included for completeness. Any reference to contact

individuals was approved. An online version with functionality to add lines is maintained at http://www.optogenetics.org/flp_lines.html. A, random inte-

gration; B, bacterial artifical chromosome (BAC) transgenic; C, start codon knockin; D, internal ribosomal entry site (IRES) knockin; E, 2a knockin; F, 2a

knockin/driver knock-out; Jackson, Jackson Laboratory (https://www.jax.org/); RBRC: Riken BioResource Research Center (https://mus.brc.riken.jp/

en/); MMRRC, Mutant Mouse Resource and Research Centers (https://www.mmrrc.org); CARD, Center for Animal Resources and Development

(http://cardb.cc.kumamoto-u.ac.jp/transgenic/index.jsp); EMMA, European Mouse Mutant Archive (https://www.infrafrontier.eu/).aTransgenic rat line.

llNeuroResource


Understanding and characterizing the limitations of these (and

any other) tools is a critical component of successful experi-

mental design and interpretation. Here we describe the function

of the intersectional toolbox using a variety of platforms in vitro

and in vivo. To assist users with designing experiments and

important controls to consider, we maintain a freely available

resource (http://www.optogenetics.org/intrsect_sop.pdf), up-

dated in response to feedback from the community, that also de-

scribes common pitfalls we and others have experienced when

scaling and implementing recombinase-based systems

(Figure 2H).

Further extending the precision with which biologists are able

to functionally interrogate cell populations is critical and

becoming more important as the field progresses. Going for-

ward, strategies such as iterative improvement of Con/Foff con-

structs, incorporation of additional molecular tools (Hafner et al.,

2019; Platt et al., 2014; Villette et al., 2019), and extension tomul-

tiple-recombinase-dependent RNA polymerase III (Ventura

et al., 2004; Yu and McMahon, 2006) expression will continue

to push boundaries of intersectional experimental design. In

this way, our rapidly advancing ability to describe cell types

through multimodal data streams integrating single-cell gene


expression with developmental (Asp et al., 2019; Petropoulos

et al., 2016), anatomical (Codeluppi et al., 2018; Satija et al.,

2015), subcellular (Chen et al., 2015), and three-dimensional

architectural (Wang et al., 2018) features at high resolution may

become accompanied by an improved capability to investigate

the functional relevance of these cell types to health or disease

in living systems.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

d KEY RESOURCES TABLE

d RESOURCE AVAILABILITY

B Lead Contact

B Materials Availability

B Data and Code Availability

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

B Animals

B Flp Line Database

B Primary Neuronal Cell Cultures

http://www.optogenetics.org/intrsect_sop.pdf









https://www.jax.org/

https://mus.brc.riken.jp/en/

https://mus.brc.riken.jp/en/

https://www.mmrrc.org

http://cardb.cc.kumamoto-u.ac.jp/transgenic/index.jsp

https://www.infrafrontier.eu/

llNeuroResource


B HEK293 Cell Cultures

d METHOD DETAILS

B Molecular cloning

B sRGECO and oScarlet development

B mRNA isolation and cDNA synthesis

B Flow cytometry

B Primary Neuron Culture Transfection

B Primary Culture Electrophysiology

B Cultured Neuron Calcium Imaging

B Virus production

B Stereotactic injections

B Slice electrophysiology

B Spectrophotometry and Viral Kinetic Analysis

B Fiber Photometry

B Histology

d QUANTIFICATION AND STATISTICAL ANALYSIS

B Additional Resources

SUPPLEMENTAL INFORMATION

Supplemental Information can be found online at https://doi.org/10.1016/j.

neuron.2020.06.003.

ACKNOWLEDGMENTS

We thank Meredith Weglarz at the Stanford FACS Facility, the Stanford

Department of Statistics, and Ethan Richman for assistance with bio-statistics

and Javier Fernandez-Alcudia and the Stanford Gene Vector and Virus Core

for virus production. S.V. received support from the NSF, Y.S.K. from Stanford

Bio-X and the Kwanjeong Fellowship, L.E.F. from NIMH and Stanford Psychi-

atry, and K.D. from the NSF, NIH, DARPA, Gatsby Foundation, NOMIS Foun-

dation, Wiegers Foundation, Tarlton Foundation, and Fresenius Foundation.

AUTHOR CONTRIBUTIONS

Conceptualization, L.E.F., C.R., Y.S.K., and K.D.; Methodology, L.E.F., C.R.,

Y.S.K., and S.V.; Software, L.E.F. and S.V.; Formal Analysis, L.E.F., Y.S.K.,

S.V., and M.I.; Experimental Investigation, L.E.F., C.R., Y.S.K., K.E.E., M.L.,

S.V., M.I., K.Y.M.C., E.Y., N.P., and A.S.O.H.; Visualization, L.E.F. and

Y.S.K.; Writing, L.E.F., C.R., Y.S.K., and K.D.; Supervision, K.D.

DECLARATION OF INTERESTS

The authors declare no competing interests.

Received: March 20, 2020

Revised: May 8, 2020

Accepted: May 29, 2020

Published: June 22, 2020

SUPPORTING CITATIONS

The following references appear in the Supplemental Information: Berndt

et al. (2016).

REFERENCES

Abraira, V.E., Kuehn, E.D., Chirila, A.M., Springel, M.W., Toliver, A.A.,

Zimmerman, A.L., Orefice, L.L., Boyle, K.A., Bai, L., Song, B.J., et al. (2017).

The cellular and synaptic architecture of the mechanosensory dorsal horn.

Cell 168, 295–310.e19.

Alhadeff, A.L., Su, Z., Hernandez, E., Klima, M.L., Phillips, S.Z., Holland, R.A.,

Guo, C., Hantman, A.W., De Jonghe, B.C., and Betley, J.N. (2018). A Neural

Circuit for the Suppression of Pain by a Competing Need State. Cell 173,

140–152.e15.

Andrews, B.J., Proteau, G.A., Beatty, L.G., and Sadowski, P.D. (1985). The

FLP recombinase of the 2micron circle DNA of yeast: interaction with its target

sequences. Cell 40, 795–803.

Aschauer, D.F., Kreuz, S., and Rumpel, S. (2013). Analysis of transduction ef-

ficiency, tropism and axonal transport of AAV serotypes 1, 2, 5, 6, 8 and 9 in the

mouse brain. PLoS ONE 8, e76310.

Asp, M., Giacomello, S., Larsson, L., Wu, C., F€urth, D., Qian, X., W€ardell, E.,

Custodio, J., Reimegard, J., Salmen, F., et al. (2019). A Spatiotemporal

Organ-Wide Gene Expression and Cell Atlas of the Developing Human

Heart. Cell 179, 1647–1660.e19.

Atasoy, D., Aponte, Y., Su, H.H., and Sternson, S.M. (2008). A FLEX switch tar-

gets Channelrhodopsin-2 to multiple cell types for imaging and long-range cir-

cuit mapping. J. Neurosci. 28, 7025–7030.

Awatramani, R., Soriano, P., Rodriguez, C., Mai, J.J., and Dymecki, S.M.

(2003). Cryptic boundaries in roof plate and choroid plexus identified by inter-

sectional gene activation. Nat. Genet. 35, 70–75.

Berndt, A., Lee, S.Y., Wietek, J., Ramakrishnan, C., Steinberg, E.E., Rashid,

A.J., Kim, H., Park, S., Santoro, A., Frankland, P.W., et al. (2016). Structural

foundations of optogenetics: Determinants of channelrhodopsin ion selec-

tivity. Proc. Natl. Acad. Sci. USA 113, 822–829.

Bindels, D.S., Haarbosch, L., van Weeren, L., Postma, M., Wiese, K.E.,

Mastop, M., Aumonier, S., Gotthard, G., Royant, A., Hink, M.A., and Gadella,

T.W., Jr. (2017). mScarlet: a bright monomeric red fluorescent protein for

cellular imaging. Nat. Methods 14, 53–56.

Bourane, S., Duan, B., Koch, S.C., Dalet, A., Britz, O., Garcia-Campmany, L.,

Kim, E., Cheng, L., Ghosh, A., Ma, Q., et al. (2015). Gate control of mechanical

itch by a subpopulation of spinal cord interneurons. Science 350, 550–554.

Britz, O., Zhang, J., Grossmann, K.S., Dyck, J., Kim, J.C., Dymecki, S.,

Gosgnach, S., and Goulding, M. (2015). A genetically defined asymmetry un-

derlies the inhibitory control of flexor–extensor locomotor movements. eLife 4,

e04718.

Brunak, S., Engelbrecht, J., and Knudsen, S. (1991). Prediction of human

mRNA donor and acceptor sites from the DNA sequence. J. Mol. Biol.

220, 49–65.

Buchholz, F., Angrand, P.O., and Stewart, A.F. (1998). Improved properties of

FLP recombinase evolved by cycling mutagenesis. Nat. Biotechnol. 16,

657–662.

Burger, C., Gorbatyuk, O.S., Velardo, M.J., Peden, C.S., Williams, P.,

Zolotukhin, S., Reier, P.J., Mandel, R.J., and Muzyczka, N. (2004).

Recombinant AAV viral vectors pseudotyped with viral capsids from serotypes

1, 2, and 5 display differential efficiency and cell tropism after delivery to

different regions of the central nervous system. Mol. Ther. 10, 302–317.

Cai, X., Kardon, A.P., Snyder, L.M., Kuzirian, M.S., Minestro, S., de Souza, L.,

Rubio, M.E., Maricich, S.M., and Ross, S.E. (2016). Bhlhb5:flpo allele uncovers

a requirement for Bhlhb5 for the development of the dorsal cochlear nucleus.

Dev. Biol. 414, 149–160.

Chakrabarty, P., Rosario, A., Cruz, P., Siemienski, Z., Ceballos-Diaz, C.,

Crosby, K., Jansen, K., Borchelt, D.R., Kim, J.-Y., Jankowsky, J.L., et al.

(2013). Capsid serotype and timing of injection determines AAV transduction

in the neonatal mice brain. PLoS ONE 8, e67680.

Chen, K.H., Boettiger, A.N., Moffitt, J.R.,Wang, S., and Zhuang, X. (2015). RNA

Imaging. Spatially resolved, highly multiplexed RNA profiling in single cells.

Science 348, aaa6090.

Chowdhury, S., Hung, C.J., Izawa, S., Inutsuka, A., Kawamura, M.,

Kawashima, T., Bito, H., Imayoshi, I., Abe, M., Sakimura, K., and Yamanaka,

A. (2019). Dissociating orexin-dependent and -independent functions of orexin

neurons using novel Orexin-Flp knock-in mice. eLife 8, e44927.

Christenson Wick, Z., Tetzlaff, M.R., and Krook-Magnuson, E. (2019). Novel

long-range inhibitory nNOS-expressing hippocampal cells. eLife 8, e46816.

Chuhma, N.,Mingote, S., Yetnikoff, L., Kalmbach, A., Ma, T., Ztaou, S., Sienna,

A., Tepler, S., Poulin, J., Ansorge, M., et al. (2018). Dopamine neuron




http://refhub.elsevier.com/S0896-6273(20)30432-3/sref1







































































llNeuroResource


glutamate cotransmission evokes a delayed excitation in lateral dorsal striatal

cholinergic interneurons. eLife 7, e39786.

Clemente-Perez, A., Ritter Makinson, S., Higashikubo, B., Brovarno, S., Cho,

F.S., Urry, A., Holden, S.S., Wimer, M., David, C., Fenno, L.E., et al. (2017).

Distinct Thalamic Reticular Cell Types Differentially Modulate Normal and

Pathological Cortical Rhythms. Cell Rep. 19, 2130–2142.

Codeluppi, S., Borm, L.E., Zeisel, A., La Manno, G., van Lunteren, J.A.,

Svensson, C.I., and Linnarsson, S. (2018). Spatial organization of the somato-

sensory cortex revealed by osmFISH. Nat. Methods 15, 932–935.

Cristinelli, S., and Ciuffi, A. (2018). The use of single-cell RNA-Seq to under-

stand virus-host interactions. Curr. Opin. Virol. 29, 39–50.

Cummings, K.A., and Clem, R.L. (2020). Prefrontal somatostatin interneurons

encode fear memory. Nat. Neurosci. 23, 61–74.

Daigle, T.L., Madisen, L., Hage, T.A., Valley, M.T., Knoblich, U., Larsen, R.S.,

Takeno, M.M., Huang, L., Gu, H., Larsen, R., et al. (2018). A Suite of

Transgenic Driver and Reporter Mouse Lines with Enhanced Brain-Cell-Type

Targeting and Functionality. Cell 174, 465–480.e22.

Dana, H., Mohar, B., Sun, Y., Narayan, S., Gordus, A., Hasseman, J.P.,

Tsegaye, G., Holt, G.T., Hu, A., Walpita, D., et al. (2016). Sensitive red protein

calcium indicators for imaging neural activity. eLife 5, e12727.

Davidson, B.L., Stein, C.S., Heth, J.A., Martins, I., Kotin, R.M., Derksen, T.A.,

Zabner, J., Ghodsi, A., and Chiorini, J.A. (2000). Recombinant adeno-associ-

ated virus type 2, 4, and 5 vectors: transduction of variant cell types and re-

gions in the mammalian central nervous system. Proc. Natl. Acad. Sci. USA

97, 3428–3432.

Diester, I., Kaufman, M.T., Mogri, M., Pashaie, R., Goo, W., Yizhar, O.,

Ramakrishnan, C., Deisseroth, K., and Shenoy, K.V. (2011). An optogenetic

toolbox designed for primates. Nat. Neurosci. 14, 387–397.

Ding, J., Luo, A.F., Hu, L., Wang, D., and Shao, F. (2014). Structural basis of the

ultrasensitive calcium indicator GCaMP6. Sci. China Life Sci. 57, 269–274.

Dymecki, S.M., and Tomasiewicz, H. (1998). Using Flp-recombinase to char-

acterize expansion of Wnt1-expressing neural progenitors in the mouse.

Dev. Biol. 201, 57–65.

Ecker, J.R., Geschwind, D.H., Kriegstein, A.R., Ngai, J., Osten, P., Polioudakis,

D., Regev, A., Sestan, N., Wickersham, I.R., and Zeng, H. (2017). The BRAIN

Initiative Cell Census Consortium: Lessons Learned toward Generating a

Comprehensive Brain Cell Atlas. Neuron 96, 542–557.

Fadok, J.P., Krabbe, S., Markovic, M., Courtin, J., Xu, C., Massi, L., Botta, P.,

Bylund, K., M€uller, C., Kovacevic, A., et al. (2017). A competitive inhibitory cir-

cuit for selection of active and passive fear responses. Nature 542, 96–100.

Fenno, L., Yizhar, O., and Deisseroth, K. (2011). The development and applica-

tion of optogenetics. Annu. Rev. Neurosci. 34, 389–412.

Fenno, L.E., Mattis, J., Ramakrishnan, C., Hyun, M., Lee, S.Y., He, M.,

Tucciarone, J., Selimbeyoglu, A., Berndt, A., Grosenick, L., et al. (2014).

Targeting cells with single vectors using multiple-feature Boolean logic. Nat.

Methods 11, 763–772.

Fenno, L.E., Mattis, J., Ramakrishnan, C., and Deisseroth, K. (2017). A guide to

creating and testing new INTRSECT constructs. Curr. Protoc. Neurosci. 80,

4.39.1–4.39.24.

Fischer, K.B., Collins, H.K., and Callaway, E.M. (2019). Sources of off-target

expression from recombinase-dependent AAV vectors and mitigation with

cross-over insensitive ATG-out vectors. Proc. Natl. Acad. Sci. U.S.A. 116,

27001–27010.

Gao, Z.R., Chen, W.Z., Liu, M.Z., Chen, X.J., Wan, L., Zhang, X.Y., Yuan, L.,

Lin, J.K., Wang, M., Zhou, L., et al. (2019). Tac1-Expressing Neurons in the

Periaqueductal Gray Facilitate the Itch-Scratching Cycle via Descending

Regulation. Neuron 101, 45–59.e9.

Gokmen, M.R., Dong, R., Kanhere, A., Powell, N., Perucha, E., Jackson, I.,

Howard, J.K., Hernandez-Fuentes, M., Jenner, R.G., and Lord, G.M. (2013).

Genome-wide regulatory analysis reveals that T-bet controls Th17 lineage dif-

ferentiation through direct suppression of IRF4. J. Immunol. 191, 5925–5932.


Goldberg, S., Adair-Kirk, T.L., Senior, R.M., and Miner, J.H. (2010).

Maintenance of glomerular filtration barrier integrity requires laminin a5.

J. Am. Soc. Nephrol. 21, 579–586.

Gradinaru, V., Zhang, F., Ramakrishnan, C., Mattis, J., Prakash, R., Diester, I.,

Goshen, I., Thompson, K.R., and Deisseroth, K. (2010). Molecular and cellular

approaches for diversifying and extending optogenetics. Cell 141, 154–165.

Gunaydin, L.A., Grosenick, L., Finkelstein, J.C., Kauvar, I.V., Fenno, L.E.,

Adhikari, A., Lammel, S., Mirzabekov, J.J., Airan, R.D., Zalocusky, K.A.,

et al. (2014). Natural neural projection dynamics underlying social behavior.

Cell 157, 1535–1551.

Hafner, G., Witte, M., Guy, J., Subhashini, N., Fenno, L.E., Ramakrishnan, C.,

Kim, Y.S., Deisseroth, K., Callaway, E.M., Oberhuber, M., et al. (2019).

Mapping Brain-Wide Afferent Inputs of Parvalbumin-Expressing GABAergic

Neurons in Barrel Cortex Reveals Local and Long-Range Circuit Motifs. Cell

Rep. 28, 3450–3461.e8.

Hara, T., and Verma, I.M. (2019). Modeling gliomas using two recombinases.

Cancer Res. 79, 3983–3991.

He, M., Tucciarone, J., Lee, S., Nigro, M.J., Kim, Y., Levine, J.M., Kelly, S.M.,

Krugikov, I., Wu, P., Chen, Y., et al. (2016). Strategies and Tools for

Combinatorial Targeting of GABAergic Neurons in Mouse Cerebral Cortex.

Neuron 91, 1228–1243.

Hirsch, M.-R., d’Autreaux, F., Dymecki, S.M., Brunet, J.-F., and Goridis, C.

(2013). A Phox2b:FLPo transgenic mouse line suitable for intersectional ge-

netics. Genesis 51, 506–514.

Hordeaux, J., Wang, Q., Katz, N., Buza, E.L., Bell, P., and Wilson, J.M. (2018).

The Neurotropic Properties of AAV-PHP.B Are Limited to C57BL/6JMice. Mol.

Ther. 26, 664–668.

Hunter, N.L., Awatramani, R.B., Farley, F.W., and Dymecki, S.M. (2005).

Ligand-activated Flpe for temporally regulated gene modifications. Genesis

41, 99–109.

Imayoshi, I., Hirano, K., Sakamoto, M., Miyoshi, G., Imura, T., Kitano, S.,

Miyachi, H., and Kageyama, R. (2012). A multifunctional teal-fluorescent

Rosa26 reporter mouse line for Cre- and Flp-mediated recombination.

Neurosci. Res. 73, 85–91.

Inoue, M., Takeuchi, A., Manita, S., Horigane, S.I., Sakamoto, M., Kawakami,

R., Yamaguchi, K., Otomo, K., Yokoyama, H., Kim, R., et al. (2019). Rational

Engineering of XCaMPs, a Multicolor GECI Suite for In Vivo Imaging of

Complex Brain Circuit Dynamics. Cell 177, 1346–1360.e24.

Jensen, P., Farago, A.F., Awatramani, R.B., Scott, M.M., Deneris, E.S., and

Dymecki, S.M. (2008). Redefining the serotonergic system by genetic lineage.

Nat. Neurosci. 11, 417–419.

Jinno, S. (2009). Structural organization of long-range GABAergic projection

system of the hippocampus. Front. Neuroanat. 3, 13.

Kanki, H., Suzuki, H., and Itohara, S. (2006). High-efficiency CAG-FLPe deleter

mice in C57BL/6J background. Exp. Anim. 55, 137–141.

Karimova, M., Splith, V., Karpinski, J., Pisabarro, M.T., and Buchholz, F.

(2016). Discovery of Nigri/nox and Panto/pox site-specific recombinase sys-

tems facilitates advanced genome engineering. Sci. Rep. 6, 30130.

Karimova, M., Baker, O., Camgoz, A., Naumann, R., Buchholz, F., and

Anastassiadis, K. (2018). A single reporter mouse line for Vika, Flp, Dre, and

Cre-recombination. Sci. Rep. 8, 14453.

Katayama, H., Yamamoto, A., Mizushima, N., Yoshimori, T., and Miyawaki, A.

(2008). GFP-like proteins stably accumulate in lysosomes. Cell Struct. Funct.

33, 1–12.

Kato, H.E., Kim, Y.S., Paggi, J.M., Evans, K.E., Allen, W.E., Richardson, C.,

Inoue, K., Ito, S., Ramakrishnan, C., Fenno, L.E., et al. (2018). Structural mech-

anisms of selectivity and gating in anion channelrhodopsins. Nature 561,

349–354.

Klein, R.L., Dayton, R.D., Leidenheimer, N.J., Jansen, K., Golde, T.E., and

Zweig, R.M. (2006). Efficient neuronal gene transfer with AAV8 leads to neuro-

toxic levels of tau or green fluorescent proteins. Mol. Ther. 13, 517–527.




























































































































llNeuroResource


Kouyama, T., Kanada, S., Takeguchi, Y., Narusawa, A., Murakami, M., and

Ihara, K. (2010). Crystal structure of the light-driven chloride pump halorho-

dopsin from Natronomonas pharaonis. J. Mol. Biol. 396, 564–579.

Lao, Z., Raju, G.P., Bai, C.B., and Joyner, A.L. (2012). MASTR: a technique for

mosaic mutant analysis with spatial and temporal control of recombination us-

ing conditional floxed alleles in mice. Cell Rep. 2, 386–396.

Lazaridis, I., Tzortzi, O., Weglage, M., M€artin, A., Xuan, Y., Parent, M.,

Johansson, Y., Fuzik, J., F€urth, D., Fenno, L.E., et al. (2019). A hypothala-

mus-habenula circuit controls aversion. Mol. Psychiatry 24, 1351–1368.

L€uond, F., Bill, R., Vettiger, A., Oller, H., Pelczar, P., and Christofori, G. (2019).

A Transgenic MMTV-Flippase Mouse Line for Molecular Engineering in

Mammary Gland and Breast Cancer Mouse Models. J. Mammary Gland

Biol. Neoplasia 24, 39–45.

Madisen, L., Zwingman, T.A., Sunkin, S.M., Oh, S.W., Zariwala, H.A., Gu, H.,

Ng, L.L., Palmiter, R.D., Hawrylycz, M.J., Jones, A.R., et al. (2010). A robust

and high-throughput Cre reporting and characterization system for the whole

mouse brain. Nat. Neurosci. 13, 133–140.

Madisen, L., Garner, A.R., Shimaoka, D., Chuong, A.S., Klapoetke, N.C., Li, L.,

van der Bourg, A., Niino, Y., Egolf, L., Monetti, C., et al. (2015). Transgenic mice

for intersectional targeting of neural sensors and effectors with high specificity

and performance. Neuron 85, 942–958.

Marcinkiewcz, C.A., Mazzone, C.M., D’Agostino, G., Halladay, L.R.,

Hardaway, J.A., DiBerto, J.F., Navarro, M., Burnham, N., Cristiano, C.,

Dorrier, C.E., et al. (2016). Serotonin engages an anxiety and fear-promoting

circuit in the extended amygdala. Nature 537, 97–101.

Matho, K.S., Huilgol, D., Galbavy, W., Kim, G., He, M., An, X., Lu, J., Wu, P., Di

Bella, D.J., Shetty, A.S., Palaniswamy, R., Hatfield, J., Raudales, R.,

Narasimhan, A., Gamache, E., Levine, J., Tucciarone, J., Mitra, P.P., Osten,

P., Arlotta, P., and Huang, Z.J. (2020). Genetic dissection of glutamatergic

neuron subpopulations and developmental trajectories in the cerebral cortex.

bioRxiv. https://doi.org/10.1101/2020.04.22.054064.

Mingote, S., Amsellem, A., Kempf, A., Rayport, S., and Chuhma, N. (2019).

Dopamine-glutamate neuron projections to the nucleus accumbens medial

shell and behavioral switching. Neurochem. Int. 129, 104482.

Miyoshi, G., Hjerling-Leffler, J., Karayannis, T., Sousa, V.H., Butt, S.J.B.,

Battiste, J., Johnson, J.E., Machold, R.P., and Fishell, G. (2010). Genetic

fate mapping reveals that the caudal ganglionic eminence produces a large

and diverse population of superficial cortical interneurons. J. Neurosci. 30,

1582–1594.

Nakano, M., Ishimura, M., Chiba, J., Kanegae, Y., and Saito, I. (2001). DNA

substrates influence the recombination efficiency mediated by FLP recombi-

nase expressed in mammalian cells. Microbiol. Immunol. 45, 657–665.

Nathanson, J.L., Yanagawa, Y., Obata, K., and Callaway, E.M. (2009).

Preferential labeling of inhibitory and excitatory cortical neurons by endoge-

nous tropism of adeno-associated virus and lentivirus vectors. Neuroscience

161, 441–450.

Petropoulos, S., Edsg€ard, D., Reinius, B., Deng, Q., Panula, S.P., Codeluppi,

S., Plaza Reyes, A., Linnarsson, S., Sandberg, R., and Lanner, F. (2016).

Single-Cell RNA-Seq Reveals Lineage and X Chromosome Dynamics in

Human Preimplantation Embryos. Cell 165, 1012–1026.

Piccirillo, R., Palmisano, I., Innamorati, G., Bagnato, P., Altimare, D., and

Schiaffino, M.V. (2006). An unconventional dileucine-based motif and a novel

cytosolic motif are required for the lysosomal and melanosomal targeting of

OA1. J. Cell Sci. 119, 2003–2014.

Platt, R.J., Chen, S., Zhou, Y., Yim, M.J., Swiech, L., Kempton, H.R., Dahlman,

J.E., Parnas, O., Eisenhaure, T.M., Jovanovic, M., et al. (2014). CRISPR-Cas9

knockin mice for genome editing and cancer modeling. Cell 159, 440–455.

Plummer, N.W., Evsyukova, I.Y., Robertson, S.D., de Marchena, J., Tucker,

C.J., and Jensen, P. (2015). Expanding the power of recombinase-based la-

beling to uncover cellular diversity. Development 142, 4385–4393.

Poulin, J.F., Caronia, G., Hofer, C., Cui, Q., Helm, B., Ramakrishnan, C., Chan,

C.S., Dombeck, D.A., Deisseroth, K., and Awatramani, R. (2018). Mapping pro-

jections ofmolecularly defined dopamine neuron subtypes using intersectional

genetic approaches. Nat. Neurosci. 21, 1260–1271.

Raymond, C.S., and Soriano, P. (2007). High-efficiency FLP and PhiC31 site-

specific recombination in mammalian cells. PLoS ONE 2, e162.

Reimsnider, S., Manfredsson, F.P., Muzyczka, N., and Mandel, R.J. (2007).

Time course of transgene expression after intrastriatal pseudotyped rAAV2/

1, rAAV2/2, rAAV2/5, and rAAV2/8 transduction in the rat. Mol. Ther. 15,

1504–1511.

Ringrose, L., Lounnas, V., Ehrlich, L., Buchholz, F., Wade, R., and Stewart, A.F.

(1998). Comparative kinetic analysis of FLP and cre recombinases: mathemat-

ical models for DNA binding and recombination. J. Mol. Biol. 284, 363–384.

Robertson, S.D., Plummer, N.W., de Marchena, J., and Jensen, P. (2013).

Developmental origins of central norepinephrine neuron diversity. Nat.

Neurosci. 16, 1016–1023.

Rodrıguez, C.I., Buchholz, F., Galloway, J., Sequerra, R., Kasper, J., Ayala, R.,

Stewart, A.F., and Dymecki, S.M. (2000). High-efficiency deleter mice show

that FLPe is an alternative to Cre-loxP. Nat. Genet. 25, 139–140.

Satija, R., Farrell, J.A., Gennert, D., Schier, A.F., and Regev, A. (2015). Spatial

reconstruction of single-cell gene expression data. Nat. Biotechnol. 33,

495–502.

Schlake, T., and Bode, J. (1994). Use of mutated FLP recognition target (FRT)

sites for the exchange of expression cassettes at defined chromosomal loci.

Biochemistry 33, 12746–12751.

Schonhuber, N., Seidler, B., Schuck, K., Veltkamp, C., Schachtler, C.,

Zukowska, M., Eser, S., Feyerabend, T.B., Paul, M.C., Eser, P., et al. (2014).

A next-generation dual-recombinase system for time- and host-specific tar-

geting of pancreatic cancer. Nat. Med. 20, 1340–1347.

Schuman, B., Machold, R.P., Hashikawa, Y., Fuzik, J., Fishell, G.J., and Rudy,

B. (2019). Four unique interneuron populations reside in neocortical layer 1.

J. Neurosci. 39, 125–139.

Sohal, V.S., Zhang, F., Yizhar, O., and Deisseroth, K. (2009). Parvalbumin neu-

rons and gamma rhythms enhance cortical circuit performance. Nature 459,

698–702.

Sun, J.J., and Ray, R. (2016). Generation of Two Noradrenergic-Specific

Dopamine-Beta-Hydroxylase-FLPo Knock-In Mice Using CRISPR/Cas9-

Mediated Targeting in Embryonic Stem Cells. PLoS ONE 11, e0159474.

Suzuki, E., and Nakayama, M. (2011). VCre/VloxP and SCre/SloxP: new site-

specific recombination systems for genome engineering. Nucleic Acids Res.

39, e49.

Takeuchi, T., Nomura, T., Tsujita, M., Suzuki, M., Fuse, T., Mori, H., and

Mishina, M. (2002). Flp recombinase transgenic mice of C57BL/6 strain for

conditional gene targeting. Biochem. Biophys. Res. Commun. 293, 953–957.

Tervo, D.G.R., Hwang, B.Y., Viswanathan, S., Gaj, T., Lavzin, M., Ritola, K.D.,

Lindo, S., Michael, S., Kuleshova, E., Ojala, D., et al. (2016). A Designer AAV

Variant Permits Efficient Retrograde Access to Projection Neurons. Neuron

92, 372–382.

Ting, J.T., Daigle, T.L., Chen, Q., and Feng, G. (2014). Acute Brain Slice

Methods for Adult and Aging Animals: Application of Targeted Patch Clamp

Analysis and Optogenetics (Humana Press), pp. 221–242.

Tovote, P., Esposito, M.S., Botta, P., Chaudun, F., Fadok, J.P., Markovic, M.,

Wolff, S.B., Ramakrishnan, C., Fenno, L., Deisseroth, K., et al. (2016). Midbrain

circuits for defensive behaviour. Nature 534, 206–212.

Ventura, A., Meissner, A., Dillon, C.P., McManus, M., Sharp, P.A., Van Parijs,

L., Jaenisch, R., and Jacks, T. (2004). Cre-lox-regulated conditional RNA inter-

ference from transgenes. Proc. Natl. Acad. Sci. USA 101, 10380–10385.

Villette, V., Chavarha, M., Dimov, I.K., Bradley, J., Pradhan, L., Mathieu, B.,

Evans, S.W., Chamberland, S., Shi, D., Yang, R., et al. (2019). Ultrafast Two-

Photon Imaging of a High-Gain Voltage Indicator in Awake Behaving Mice.

Cell 179, 1590–1608.e23.

Vooijs, M., van der Valk, M., te Riele, H., and Berns, A. (1998). Flp-mediated

tissue-specific inactivation of the retinoblastoma tumor suppressor gene in

the mouse. Oncogene 17, 1–12.






























https://doi.org/10.1101/2020.04.22.054064































































































llNeuroResource


Vrontou, S., Wong, A.M., Rau, K.K., Koerber, H.R., and Anderson, D.J. (2013).

Genetic identification of C fibres that detect massage-like stroking of hairy skin

in vivo. Nature 493, 669–673.

Wang, X., Allen, W.E., Wright, M.A., Sylwestrak, E.L., Samusik, N., Vesuna, S.,

Evans, K., Liu, C., Ramakrishnan, C., Liu, J., et al. (2018). Three-dimensional

intact-tissue sequencing of single-cell transcriptional states. Science 361,

eaat5691.

Weinberg, B.H., Pham, N.T.H., Caraballo, L.D., Lozanoski, T., Engel, A.,

Bhatia, S., and Wong, W.W. (2017). Large-scale design of robust genetic cir-

cuits with multiple inputs and outputs for mammalian cells. Nat. Biotechnol.

35, 453–462.

Wojcinski, A., Morabito, M., Lawton, A.K., Stephen, D.N., and Joyner, A.L.

(2019). Genetic deletion of genes in the cerebellar rhombic lip lineage can stim-

ulate compensation through adaptive reprogramming of ventricular zone-

derived progenitors. Neural Dev. 14, 4.

Wu, Y., Wang, C., Sun, H., LeRoith, D., and Yakar, S. (2009). High-efficient

FLPo deleter mice in C57BL/6J background. PLoS ONE 4, e8054.

Wu, J., Liu, X., Nayak, S.G., Pitarresi, J.R., Cuitino, M.C., Yu, L., Hildreth, B.E.,

3rd, Thies, K.A., Schilling, D.J., Fernandez, S.A., et al. (2017). Generation of a

pancreatic cancer model using a Pdx1-Flp recombinase knock-in allele. PLoS

ONE 12, e0184984.


Yizhar, O., Fenno, L.E., Davidson, T.J., Mogri, M., and Deisseroth, K. (2011).

Optogenetics in neural systems. Neuron 71, 9–34.

Yu, J., and McMahon, A.P. (2006). Reproducible and inducible knockdown of

gene expression in mice. Genesis 44, 252–261.

Yu, K., Ahrens, S., Zhang, X., Schiff, H., Ramakrishnan, C., Fenno, L.,

Deisseroth, K., Zhao, F., Luo, M.H., Gong, L., et al. (2017). The central amyg-

dala controls learning in the lateral amygdala. Nat. Neurosci. 20, 1680–1685.

Yu, J.Y., Pettibone, J.R., Guo, C., Zhang, S., Saunders, T.L., Hughes, E.D.,

Filipiak, W.E., Zeidler, M.G., Bender, K.J., Hopf, F., et al. (2018). Knock-in

rats expressing Cre and Flp recombinases at the Parvalbumin locus.

BioRxiv. https://doi.org/10.1101/386474.

Zhang, F. (2008). Larry M. Katz Memorial Lecture, Cold Spring Harbor

Laboratory Meeting on Neuronal Circuits (Cold Spring Harbor Laboratory,

March 2008).

Zhang, M.D., Su, J., Adori, C., Cinquina, V., Malenczyk, K., Girach, F., Peng,

C., Ernfors, P., Low, P., Borgius, L., et al. (2018). Ca2+-binding protein

NECAB2 facilitates inflammatory pain hypersensitivity. J. Clin. Invest. 128,

3757–3768.





























https://doi.org/10.1101/386474








llNeuroResource


STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER

Bacterial and Virus Strains

One Shot Stbl3 e-coli ThermoFisher Cat# C737303

Adeno-Associated Virus Coat Protein 2/8 Stanford GVVC AAV8

Adeno-Associated Virus Coat Protein 2/retro Stanford GVVC AAV-retro

pAAV-EF1a-FlpO-p2a-Cre This Paper Stanford GVVC AAV8

pAAV-EF1a-FlpO Fenno et al., 2014 Stanford GVVC AAV8

pAAV-EF1a-FlpO Fenno et al., 2014 Stanford GVVC AAV-retro

pAAV-EF1a-Cre Fenno et al., 2014 Stanford GVVC AAV8

pAAV-EF1a-VCre Fenno et al., 2014 Stanford GVVC AAV8

pAAV-EF1a-VCre Fenno et al., 2014 Stanford GVVC AAV-retro

pAAV-cDIO-EYFP Fenno et al., 2014 Stanford GVVC AAV8

pAAV-fDIO-EYFP Fenno et al., 2014 Stanford GVVC AAV8

pAAV-vcDIO-EYFP Fenno et al., 2014 Stanford GVVC AAV8

pAAV-Ef1a-EYFP This paper Stanford GVVC AAV8

pAAV-Ef1a-Con/Fon-EYFP Fenno et al., 2014 Stanford GVVC AAV8

pAAV-Ef1a-Con/Foff-EYFP Fenno et al., 2014 Stanford GVVC AAV8

pAAV-Ef1a-Coff/Fon-EYFP Fenno et al., 2014 Stanford GVVC AAV8

pAAV-Ef1a-Con/Foff 2.0-EYFP This Paper Stanford GVVC AAV8

pAAV-nEF-3x-EYFP This Paper Addgene Cat# 137163

pAAV-Ef1a-3x-GCaMP6M This Paper Addgene Cat# 137164

Chemicals, Peptides, and Recombinant Proteins

AP-V Tocris Cat# 0106

CNQX Tocris Cat# 0190

Fluorodeoxyuridine (FUDR) Sigma Cat# F0503

Lipofectamine 2000 ThermoFisher Cat# 11668027

Propidium Iodide Sigma Cat# P4170

RNeasy Mini Kit QIAGEN Cat # 74106

SuperScript III One-Step RT-PCR System ThermoFisher Cat # 12574026

Tetrodotoxin Tocris Cat #1078

Experimental Models: Cell Lines

Rat Hippocampal Primary Neurons Derived In-House N/A

HEK293 Cells ThermoFisher Scientific Cat # R70007

Experimental Models: Organisms/Strains

WT C57/BL6 mice Jackson Laboratories Cat# 000664

SST-Cre C57/BL6 transgenic mice Jackson Laboratories Cat# 013044

TH-Cre C57/BL6 transgenic mice EMMA Cat# 00254

Sprague-Dawley rat pups Charles River n/a

Oligonucleotides

See Table S1 N/A N/A

Recombinant DNA

pAAV-EF1a-Con/Fon-GCaMP6M This Paper Addgene Cat# 137119

pAAV-Ef1a-Con/Foff-GCaMP6M This Paper N/A

pAAV-Ef1a-Con/Foff 2.0-GCaMP6M This Paper Addgene Cat# 137120

pAAV-Ef1a-Coff/Fon-GCaMP6M This Paper Addgene Cat# 137121


Neuron 107, 1–18.e1–e11, September 9, 2020 e1

Continued


pAAV-Ef1a-Con/Fon-GCaMP6F This Paper Addgene Cat# 137122

pAAV-Ef1a-Con/Foff-GCaMP6F This Paper N/A

pAAV-Ef1a-Con/Foff 2.0-GCaMP6F This Paper Addgene Cat# 137123

pAAV-Ef1a-Coff/Fon-GCaMP6F This Paper Addgene Cat# 137124

pAAV-Ef1a-sRGECO This Paper Addgene Cat# 137125

pAAV-Ef1a-Con/Fon-sRGECO This Paper Addgene Cat# 137126

pAAV-Ef1a-Con/Foff-sRGECO This Paper N/A

pAAV-Ef1a-Con/Foff 2.0-sRGECO This Paper Addgene Cat# 137127

pAAV-Ef1a-Coff/Fon-sRGECO This Paper Addgene Cat# 137128

pAAV-Ef1a-Con/Fon-BFP This Paper Addgene Cat# 137129

pAAV-Ef1a-Con/Foff-BFP This Paper N/A

pAAV-Ef1a-Con/Foff 2.0-BFP This Paper Addgene Cat# 137130

pAAV-Ef1a-Coff/Fon-BFP This Paper Addgene Cat# 137131

pAAV-Ef1a-Con/Fon-mCherry This Paper Addgene Cat# 137132

pAAV-Ef1a-Con/Foff-mCherry This Paper N/A

pAAV-Ef1a-Con/Foff 2.0-mCherry This Paper Addgene Cat# 137133

pAAV-Ef1a-Coff/Fon-mCherry This Paper Addgene Cat# 137134

pAAV-Ef1a-oScarlet This Paper Addgene Cat# 137135

pAAV-Ef1a-Con/Fon-oScarlet This Paper Addgene Cat# 137136

pAAV-Ef1a-Con/Foff-oScarlet This Paper N/A

pAAV-Ef1a-Con/Foff 2.0-oScarlet This Paper Addgene Cat# 137137

pAAV-Ef1a-Coff/Fon-oScarlet This Paper Addgene Cat# 137138

pAAV-nEF-Con/Fon-ChR2(ET/TC)-EYFP This Paper Addgene Cat# 137139

pAAV-nEF-Con/Foff-ChR2(ET/TC)-EYFP This Paper N/A

pAAV-nEF-Con/Foff 2.0-ChR2(ET/TC)-EYFP This Paper Addgene Cat# 137140

pAAV-nEF-Coff/Fon-ChR2(ET/TC)-EYFP This Paper Addgene Cat# 137141

pAAV-nEF-Con/Fon-ChR2-mCherry This Paper Addgene Cat# 137142

pAAV-nEF-Con/Foff-ChR2-mCherry This Paper N/A

pAAV-nEF-Con/Foff 2.0-ChR2-mCherry This Paper Addgene Cat# 137143

pAAV-nEF-Coff/Fon-ChR2-mCherry This Paper Addgene Cat# 137144

pAAV-nEF-Con/Fon-bREACHes-EYFP This Paper Addgene Cat# 137145

pAAV-nEF-Con/Foff-bREACHes-EYFP This Paper N/A

pAAV-nEF-Con/Foff 2.0-bREACHes-EYFP This Paper Addgene Cat# 137146

pAAV-nEF-Coff/Fon-bREACHes-EYFP This Paper Addgene Cat# 137147

pAAV-nEF-Con/Fon-Arch 3.3-p2a-EYFP This Paper Addgene Cat# 137148

pAAV-nEF-Con/Foff-Arch 3.3-EYFP This Paper N/A

pAAV-nEF-Con/Foff 2.0-Arch 3.3-EYFP This Paper Addgene Cat# 137149

pAAV-nEF-Coff/Fon-Arch 3.3-p2a-EYFP This Paper Addgene Cat# 137150

pAAV-nEF-NpHR3.3-EYFP This Paper Addgene Cat# 137151

pAAV-nEF-Con/Fon-NpHR 3.3-EYFP This Paper Addgene Cat# 137152

pAAV-nEF-Con/Foff-NpHR 3.3-EYFP This Paper N/A

pAAV-nEF-Con/Foff 2.0-NpHR 3.3-EYFP This Paper Addgene Cat# 137153

pAAV-nEF-Coff/Fon-NpHR 3.3-EYFP This Paper Addgene Cat# 137154

pAAV-nEF-Con/Fon-iC++-EYFP This Paper Addgene Cat# 137155

pAAV-nEF-Con/Foff-iC++-EYFP This Paper N/A

pAAV-nEF-Con/Foff 2.0-iC++-EYFP This Paper Addgene Cat# 137156

pAAV-nEF-Coff/Fon-iC++-EYFP This Paper Addgene Cat# 137157

pAAV-nEF-ChRmine-mScarlet This Paper Addgene Cat# 137158


llNeuroResource

e2 Neuron 107, 1–18.e1–e11, September 9, 2020


Continued


pAAV-nEF-Con/Fon-ChRmine 3.3-oScarlet This Paper Addgene Cat# 137159

pAAV-nEF-Coff/Fon-ChRmine 3.3-oScarlet This Paper Addgene Cat# 137160

pAAV-nEF-Con/Foff-ChRmine 3.3-oScarlet This Paper N/A

pAAV-nEF-Con/Foff 2.0-ChRmine 3.3-oScarlet This Paper Addgene Cat# 137161

pAAV-Ef1a-Con/Foff 2.0-EYFP This Paper Addgene Cat# 137162

pAAV-nEF-Con/Foff 2.0-ChR2-EYFP This Paper Addgene Cat# 137163

pAAV-nEF-3x-EYFP This Paper Addgene Cat# 137164

pAAV-Ef1a-3x-GCaMP6M This Paper Addgene Cat# 137165

Software and Algorithms

ImageJ NIH https://imagej.nih.gov/ij/

SnapGene GSL Biotech LLC https://www.snapgene.com:443/

FlowJo Flowjo https://www.flowjo.com

Pymol Schrodinger https://www.pymol.org/2/

Netgene2 splicing algorithm DTU Bioinformatics http://www.cbs.dtu.dk/services/NetGene2/

MATLAB MathWorks https://www.mathworks.com/products/matlab.html

Metamorph Molecular Devices https://www.moleculardevices.com

pClamp 10.6 Molecular Devices https://www.moleculardevices.com

CueMol: Molecular Visualization Framework Cuemol http://www.cuemol.org/

Other

MultiClamp700B Amplifier Molecular Devices https://www.moleculardevices.com

DigiData 1440A Molecular Devices https://www.moleculardevices.com

DM-LFSA Leica N/A

Power meter Thorlabs PM100D

200 um implants Doric MFC_200/245-0.53_3.5mm_MF2.5_FLT

400 um implants Doric MFC_400/430-0.66_15mm_MF2.5_FLT (cut to length)

DxP FACSCAN Becton Dickinson https://www.bdbiosciences.com/en-us

SPECTRA-X Light Engine Lumencor https://lumencor.com

3D-printed well insert This Paper https://www.protolabs.com

Compact CCD Spectrometer Thorlabs CCS100

Fiber Bundle Thorlabs BFL200LS02

Reflective Collimator Thorlabs RC04SMA-P01

Zoom Lens Tube Thorlabs SM1NR05

Lens Edmund Optics #62-561

Filter Holder Thorlabs CFH2

Lens Tube Coupler Thorlabs Sm1T20

Kinematic Dichroic Filter Mount, 30mm Thorlabs DFM1

End Cap Thorlabs SM1CP2

Sm1 Threaded Adaptor Thorlabs AD12F

Fixed Focus Collimator Thorlabs F240FC-A

SM1 coupler Thorlabs SM1NT

Multimode Fiber Optic Patch Cable (FCPC to SMA) Thorlabs m76L01

LED Controller Thorlabs LEDD1B

Fiber-Coupled LED Thorlabs M505F1

Excitation Bandpass Filter Thorlabs MF497-16

Dichroic Chroma T525LPXR

Emission Bandpass Filter Thorlabs MF535-22

llNeuroResource



https://imagej.nih.gov/ij/

https://www.snapgene.com:443/

https://www.flowjo.com

https://www.pymol.org/2/

http://www.cbs.dtu.dk/services/NetGene2/

https://www.mathworks.com/products/matlab.html

https://www.moleculardevices.com


http://www.cuemol.org/



https://www.bdbiosciences.com/en-us

https://lumencor.com

https://www.protolabs.com

llNeuroResource


RESOURCE AVAILABILITY

Lead ContactAll reagents detailed in this manuscript are freely available for academic use from the Deisseroth Lab (http://www.stanford.edu/

group/dlab/optogenetics/sequence_info.html) and via the Lead Contact (Karl Deisseroth; [email protected]).

Materials AvailabilityMost plasmids and viruses are also available through AddGene (http://www.addgene.org/; see Key Resources Table) and/or the

Stanford Viral Vector Core (https://neuroscience.stanford.edu/research/programs/community-labs/neuroscience-gene-vector-

and-virus-core). The Flp mouse line database is maintained by the Deisseroth Lab (http://www.stanford.edu/group/dlab/

optogenetics/flp_lines.html). A standard operating procedure for INTRSECT reagents is also available (https://web.stanford.edu/

group/dlab/optogenetics/intrsect_sop.pdf).

Data and Code AvailabilityNo stand-alone code was generated for this project, however, MATLAB code available upon request from lead contact.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

AnimalsAdult (at least 8 weeks of age) wild-type female and transgenic Ssttm(Cre) (somatostatin-IRES-Cre; Jax 013044; RRID:IMSR_-

JAX:013044) mice were group housed up to four to a cage and kept on a reverse 12-h light/dark cycle with ad libitum food and water.

Experimental protocols were approved by and conform to Stanford University IACUC and meet guidelines of the US National Insti-

tutes of Health guide for the Care and Use of Laboratory Animals. See Key Resources Table for specific transgenic animal strain

information.

Flp Line DatabaseTo locate transgenic, Flp-expressing mouse lines, the following websites were searched using the term ‘Flp’ (with the exception of

Google Scholar, where the term ‘Flp driver mouse line’ was used) in September, 2018 and November, 2019, and the results were

manually assessed.

Jackson Labs https://www.jax.org/mouse-search

MMRRC https://www.mmrrc.org/catalog/

StrainCatalogSearchForm.php

APF https://pb.apf.edu.au/phenbank/findstrains.html

NIH Blueprint http://www.credrivermice.org/

GENSAT http://www.gensat.org/cre.jsp

Taconic https://www.taconic.com/find-your-model/

Mousebook https://www.mousebook.org/

MGI http://www.informatics.jax.org/downloads/reports/

MGI_Recombinase_Full.html

RIKEN https://mus.list.brc.riken.jp/en/

RRRC http://www.rrrc.us/

Google https://scholar.google.com/

Pubmed https://pubmed.ncbi.nlm.nih.gov/

The initial publications describing the production of the lineswere assessed for the constructionmethod, promoter, and Flp variant.

In cases where animals located in publications were not found in available repositories or when animal lines in repositories were not

found in publications, the corresponding author or listed depositor was directly contacted for more information and for direction in

how other researchers can acquire the lines.

Primary Neuronal Cell CulturesPrimary cultured hippocampal neurons were prepared from P0 Sprague-Dawley rat pups (Charles River). CA1 and CA3 from entire

litters (male and female pups) were isolated, digested with 0.4 mg ml-1 papain (Worthington), and plated onto glass coverslips


http://www.stanford.edu/group/dlab/optogenetics/sequence_info.html



http://www.addgene.org/

https://neuroscience.stanford.edu/research/programs/community-labs/neuroscience-gene-vector-and-virus-core


http://www.stanford.edu/group/dlab/optogenetics/flp_lines.html


https://web.stanford.edu/group/dlab/optogenetics/intrsect_sop.pdf


https://www.jax.org/mouse-search

https://www.mmrrc.org/catalog/StrainCatalogSearchForm.php

https://www.mmrrc.org/catalog/StrainCatalogSearchForm.php

https://pb.apf.edu.au/phenbank/findstrains.html

http://www.credrivermice.org/

http://www.gensat.org/cre.jsp

https://www.taconic.com/find-your-model/

https://www.mousebook.org/

http://www.informatics.jax.org/downloads/reports/MGI_Recombinase_Full.html

http://www.informatics.jax.org/downloads/reports/MGI_Recombinase_Full.html

https://mus.list.brc.riken.jp/en/

http://www.rrrc.us/

https://scholar.google.com/

https://pubmed.ncbi.nlm.nih.gov/

llNeuroResource


precoatedwith 1:30Matrigel (BectonDickinson Labware). Cultures weremaintained in a 5%CO2 humid incubator with Neurobasal-A

media (Thermo Fisher) containing 1.25% FBS (HyClone), 4% B-27 supplement (GIBCO), 2 mMGlutamax (GIBCO) and 2 mg/ml fluo-

rodeoxyuridine (FUDR, Sigma), and grown on coverslips in a 24-well plate at a density of 65,000 cells per well.

HEK293 Cell CulturesHEK293FT cells (Thermo Fisher) were maintained in a 5%CO2 humid incubator with DMEMmedia (GIBCO) supplemented with 10%

FBS (Invitrogen), 1% Glutamax (Invitrogen), and 1x Penicillin-Streptomycin (Invitrogen). They were enzymatically passaged at 90%

confluence by trypsinization.

METHOD DETAILS

Molecular cloningAll single recombinase-dependent plasmids were constructed in an AAV-Ef1a backbone using the double-floxed inverted open-

reading-frame (DIO) strategy described previously (Atasoy et al., 2008; Sohal et al., 2009); briefly, the ORF is in the reverse comple-

ment orientation between two cassettes of recombinase recognition sites. See the Key Resources Table for information specific to

the Cre-dependent (cDIO), Flp-dependent (fDIO) vCre-dependent (vcDIO), Dre-dependent (dDIO), and sCre-dependent (scDIO) iter-

ations. dDIO rox cassette was previously described (Fenno et al., 2014). This, and all constructs used in this paper, contain the wood-

chuck hepatitis virus post-transcriptional regulatory element (WPRE) to enhance expression.

A series of recombinase-expression plasmids were used, as detailed in the Key Resources Table. All recombinase expression

plasmids were constructed in an AAV-EF1a backbone. These were made as fluorophore-expressing, bicistronic plamids (using

an internal ribosomal entry site or p2a sequence), or without any visual marker, as indicated in the text and figures. FlpO (Raymond

and Soriano, 2007) was used for Flp-dependent expression.

Mutations to NpHR, jRGECO1a, and mScarlett were introduced via overlapping PCR with primers containing the mutation. The

NpHR W179F mutation was chosen from amino acids more than 8 A away from the retinal binding pocket / ion-pumping pathway

by analyzing the crystal structure of NpHR (PDB ID: 3A7K; Kouyama et al., 2010). The jRGECO1a (Dana et al., 2016) mutation

E217D and the mScarlett (Bindels et al., 2017) mutation E95D was chosen to improve functional expression by interrupting a lyso-

somal import motif (Piccirillo et al., 2006). jRGECO1a, sRGECO, mScarlet, and oScarlet inclusion quantification was performed by

viral infection (jRGECO1a and sRGECO) or calcium-phosphate transfection (mScarlet and oScarlet) of primary neuron cultures, fol-

lowed by 4% PFA fixation, mounting, imaging on a confocal, and blinded manual analysis with randomly shuffled and anonymized

image labeling.

Tools noted to be ‘3.30 versions (NpHR, ChRmine, and Arch) include the addition of endoplasmic reticulum and membrane traf-

ficking motifs previously described (Gradinaru et al., 2010), as well as the addition of a p2a bicistronic expression sequence that al-

lows for independent translation of opsin and fluorophore from a single mRNA transcript.

INTRSECT constructs were produced with either one intron (for single ORF, two-recombinase-dependent tools), two introns (for

double ORF, two-recombinase-dependent tools), or two introns (for single ORF, three-recombinase-dependent tools).

Single ORF, two-recombinase-dependent INTRSECT plasmids were constructed by splitting the reading frame into two pieces

(referred to as ‘exon 1’ and ‘exon 20) after being analyzed for naturally occurring splice-site-like sequences using a public bioinfor-

matics tool (http://www.cbs.dtu.dk/services/NetGene2/; Brunak et al., 1991), designed as possible to have the splice site disrupt the

reading frame to decrease the chance of partial protein synthesis from exon 2 in the absence of the pre-defined expression criteria. A

derivative of the mouse IgE intron 3 (Fenno et al., 2014) containing a cDIO cassette and fDIO cassette was inserted between the

exons, with the donor and acceptor sites fused directly to the 30 terminus of exon 1 and 50 terminus of exon 2, respectively. A separate

cDIO cassette was added directly after the promoter (50 to the entire coding sequence) and a fDIO cassette was added directly before

theWPRE (30 to the entire coding sequence). To create Con/Fon constructs, both exons are in the reverse complement orientation. To

create Con/Foff constructs, exon 1 is in the reverse complement orientation and exon 2 is in the sense direction. To create Coff/Fon

constructs, exon 1 is in the sense direction and exon 2 is in the reverse complement orientation. All of these plasmids are constructed

in an AAV-EF1a backbone with a 30 WPRE and are detailed in the Key Resources Table.

Double ORF, two-recombinase-dependent INTRSECT plasmids were constructed by splitting the reading frame into three pieces

(referred to as ‘exon 1’, ‘exon 20, and ‘exon 30) after being analyzed for naturally occurring splice-site-like sequences using a public

bioinformatics tool (http://www.cbs.dtu.dk/services/NetGene2/; Brunak et al., 1991) and choosing a splice site in each of the reading

frames, designed as possible to have the splice sites disrupt the reading frame to decrease the chance of partial protein synthesis

from exon 2 or exon 3 in the absence of the pre-defined expression criteria. A derivative of the CMV Towne Variant intron B (Fenno

et al., 2014) containing a cDIO cassette was inserted between exon 1 and exon 2, with the donor and acceptor sites fused directly to

the 30 terminus of exon 1 and 50 terminus of exon 2, respectively. A derivative of the mouse IgE intron 3 (Fenno et al., 2014) containing

a cDIO cassette was inserted between exon 2 and exon 3, with the donor and acceptor sites fused directly to the 30 terminus of exon 2

and 50 terminus of exon 3, respectively. Separate fDIO cassettes were added directly after the promoter (50 to the entire coding

sequence) and directly before the WPRE (30 to the entire coding sequence). To create Con/Fon constructs, the exon order is exon

3, exon 2, exon 1, with exons 1 and 3 in the reverse complement orientation and exon 2 in the sense orientation. To create Con/

Foff constructs, the exon order is exon 1, exon 2, exon 3, with exons 1 and 3 in the sense orientation and exon 2 in the reverse




llNeuroResource


complement orientation. To create Coff/Fon constructs, the exon order is exon 3, exon 2, exon 1, with all exons in the reverse com-

plement orientation. All of these plasmids are constructed in an AAV-nEF backbone with a 30 WPRE and are detailed in the Key Re-

sources Table.

Single ORF, three-recombinase-dependent INTRSECT plasmids were constructed by splitting the reading frame into three pieces

(referred to as ‘exon 1’, ‘exon 20, and ‘exon 30) after being analyzed for naturally occurring splice-site-like sequences using a public

bioinformatics tool (http://www.cbs.dtu.dk/services/NetGene2/; Brunak et al., 1991), designed as possible to have the splice site

disrupt the reading frame to decrease the chance of partial protein synthesis from exon 2 or exon 3 in the absence of the pre-defined

expression criteria. A derivative of the CMV Towne Variant intron B (Fenno et al., 2014) containing a cDIO cassette and fDIO (F3/F5-

based) cassette was inserted between exon 1 and exon 2, with the donor and acceptor sites fused directly to the 30 terminus of exon 1

and 50 terminus of exon 2, respectively. A derivative of the mouse IgE intron 3 (Fenno et al., 2014) containing a fDIO cassette and

vcDIO cassette was inserted between exon 2 and exon 3, with the donor and acceptor sites fused directly to the 30 terminus of

exon 2 and 50 terminus of exon 3, respectively. A separate cDIO cassette was added directly after the promoter (50 to the entire coding

sequence) and a vcDIO cassette was added directly before the WPRE (30 to the entire coding sequence). To create Con/Fon/VCon

constructs, the exon order is exon 1, exon 2, exon 3, with all exons in the reverse complement orientation. All of these plasmids are

constructed in an AAV-nEF backbone with a 30 WPRE and are detailed in the Key Resources Table.

To produce FRT variants for screening Con/Foff improvements, the FRT, F3, F5 (Schlake and Bode, 1994), and F72 (Nakano et al.,

2001) sequences were built into various combinations of AAV-EF1a-Con/Foff-EYFP. In addition, a 14bp addition noted from the

genomic FRT sequence (Andrews et al., 1985) was added either to the 50 or 30 (or both) ends of F3, F5, and/or FRT in configurations.

These were synthesized de novo and incorporated into the Flp cassette using standard cloning techniques. After screening, the FRT-

F5 cassette was incorporated into all Con/Foff constructs.

FRT variant 50 repeat 8bp central motif 30 repeat

FRT gaagttcctattc tctagaaa gtataggaacttc

F3 gaagttcctattc ttcaaata gtataggaacttc

F5 gaagttcctattc ttcaaaag gtataggaacttc

F72 gaagttcctattc tgtagaaa gtataggaacttc

14bp addition gaagttcctattcc

All AAV vectors were tested for in vitro expression before viral production. Maps are available at http://www.stanford.edu/group/

dlab/optogenetics.

sRGECO and oScarlet developmentPrior to entering mScarlet and jRGECO1a into the INTRSECT production pipeline, we observed fluorescent puncta (assumed to be

protein aggregates) when expressed either in vitro (mScarlet; Figure S1B) or in vivo (jRGECO1a; Figure S2B). Both are based on

monomeric Red Fluorescent Protein (Bindels et al., 2017; Dana et al., 2016), which is known to be degradation-resistant and accu-

mulate in lysosomes (Katayama et al., 2008). We found that the unconventional lysosomal targeting motif tryptophan-glutamic acid

(Piccirillo et al., 2006) was conserved in both of these tools. To overcome this aggregation problem we mutated this motif mScar-

let(E95D) (‘oScarlet’ for ‘optimized’; Figure S1A) and jRGECO1a(E217D) (‘sRGECO’ for ‘Stanford’ and ‘smooth’; Figure S2A). oScarlet

had significantly fewer aggregates in cultured neurons (25.9 per neuron in mScarlet, 0.58 per neuron in oScarlet, p < 0.0001, Mann-

Whitney test), with no obvious difference in fluorescence intensity as assayed by flow cytometry (Figures S1B and S1C). The distri-

bution of aggregates in sRGECO were significantly less relative to jRGECO1a in mouse mPFC (p = 0.0368, Kolmogorov-Smirnov

test), without a difference in either mean aggregates per neuron or total fluorescence across the imaged cortical slices (Figures

S2B and S2C; mean aggregates 6.96 per neuron in jRGECO1a versus 5.73 per neuron in sRGECO, p = 0.7846, Mann-Whitney

test, total integrated fluorescence p = 0.1867, unpaired t test. N as indicated). To characterize this improved GECI more thoroughly,

we designed a small, 3D-printable well insert for electrical stimulation of cultured neurons using a 100 mA source applied through

platinum wires (Figure S2D). In vitro characterization of sRGECO and jRGECO1a (Figure S2E) revealed a difference in basal fluores-

cence and related differences in signal magnitude, but intact function. Considering the improved expression patterns of oScarlet and

sRGECO, we used these versions as the sequence base for creating their INTRSECT variants.

mRNA isolation and cDNA synthesisHEK293FT cells at 50% confluence were transfected with endotoxin-free DNA using Lipofectamine 2000 (Thermo Fisher) following

the manufacturer protocol. Five days post transfection, RNA extraction was performed using reagents from the RNeasy Mini Kit

(QIAGEN). Cells were disrupted with lysis buffer and homogenized using QiaShredder homogenizer columns. Combined first-strand

cDNA/PCR using the SuperScript III One-Step RT-PCR System (Thermo Fisher) was performed with the following initial reaction



http://www.stanford.edu/group/dlab/optogenetics

http://www.stanford.edu/group/dlab/optogenetics

llNeuroResource


conditions: 45�C . 30 min, 94�C . 2 min, 40 cycles of 94�C . 15 s, 45�C . 30 s, 68�C . 1 min, ending with 68�C . 5 min using various

combinations of primers (F, forward; R, reverse) as noted in the Key Resources Table and sometimes optimized further to reduce

off-target amplicons. The PCR product was electrophoresed on a 0.8% agarose gel, photographed, and DNA bands purified

from the gel and sequenced to determine splice junctions.

Flow cytometryHEK293FT cells (Thermo Fisher) were grown in 24-well tissue culture plates to 50% confluence and transfected in duplicate with

800 ng total DNA with Lipofectamine 2000 (Thermo Fisher) following the manufacturer protocol. In dual-recombinase INTRSECT

plasmid experiments, 267ng of each plasmid was used, except in recombinase ratio experiments, where a fixed amount of IN-

TRSECT plasmid and variable ratios of Cre and Flp, as noted in the figure, were added. In Triplesect experiments, 200ng of each

plasmid was added. Five days post transfection, cells were removed by enzymatic dissociation (TrypLE, GIBCO), resuspended in

cold PBS, pelleted at 200g for 5 min and resuspended in 500 mL PBS supplemented with 1 mg/mL propidium iodide (PI; Sigma,

used with green and blue fluorescent constructs) or 5 mM 4’6-diamidino-2-phenylindole (DAPI; Thermo Fisher, used with red fluores-

cent constructs), and then placed on ice under aluminum foil until analysis. Flow cytometry was completed on a DxP FACSCAN

analyzer at the Stanford Shared FACS Facility using settings optimized for side scatter (SS), forward scatter (FS), vital dye (PI or

DAPI) and fluorophore (BFP, EYFP, GCaMP6, mCherry, sRGECO, or oScarlet) acquisition using positive (non-recombinase-depen-

dent parent construct), negative (empty transfection) and dead (heat-killed; 95�C for 3 min) conditions as controls. Live-cell popula-

tions used in comparisons were isolated from debris and dead cells in post hoc analysis using FlowJo 10.4.2 (FlowJo) by (i) positively

gating for the high-density population in plotting FS versus SS and (ii) negatively gating for vital dye+ cells. Analysis and plots were

completed using FlowJo software.

Analysis of FRT cassette variants was completed by exporting the live-cell populations defined as above, then isolating the pop-

ulation of cells with EYFP expression higher than the maximum expression value of the negative population (‘residual population’).

Outliers were removed from the negative sample prior to determining the maximum fluorescence value by excluding samples with

fluorescence values greater than three median absolute deviations from the median fluorescence. We then calculated the (i) mean

fluorescence of the residual population and (ii) the percentage of the total cells represented by the residual population. Flp titration

experiments were completed by transfecting a fixed amount of Con/Foff-EYFP variant, Cre, and a varying amount of Flp to create

molar ratios as indicated in the figure. A constant amount of DNA was transfected in each condition, with the difference between

conditions made up with empty vector.

Primary Neuron Culture Transfection2ug total DNA of mixture of recombinase and INTRSECT construct, or an equivalent amount of the parental tool expression

construct, was mixed with 1.875 mL 2 M CaCl2 (final Ca2+ concentration 250 mM) in 15 mL H2O. To DNA-CaCl2 we added 15 mL

of 2 3 HEPES-buffered saline (pH 7.05). After 20 min at room temperature (20–22�C), the mix was added dropwise into each well

(from which the growth medium had been removed and replaced with pre-warmed minimal essential medium (MEM)) and transfec-

tion proceeded for 45–60min at 37�C, after which eachwell waswashedwith 33 1mLwarmMEMbefore the original growthmedium

was returned. Neurons were allowed to express transfected DNA for 6-8 days prior to experimentation.

Primary Culture ElectrophysiologyRecordings of neurons prepared and transfected as above were obtained in Tyrode’s medium (in mM: 150 NaCl, 4 KCl, 2 MgCl2, 2

CaCl2, 10 d-glucose, 10 HEPES, adjusted to pH 7.3-7.4 with NaOH, 320-330 osmolarity) supplemented with Tetrodotoxin (TTX; Toc-

ris, 1 mM), (2R)-2-amino-5-phosphonopentanoic acid (APV; Tocris, 10 mM), and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; Tocris,

25 mM) as indicated, with a standard internal solution (in mM: 130 K-gluconate, 10 KCl, 10 HEPES, 10 EGTA, 2MgCl2, to pH 7.25 with

KOH, 300-310 osmolarity) in 3- to 6-MU glass pipettes. Light from a SPECTRA-X Light Engine (Lumencor) with LEDs with individual

light power adjusted for uniform light power density of 1mW/mm2 across wavelengths and 470/15, 513/15, and 585/29 filters were

used for blue, green and orange illumination, respectively. The Spectra X was coupled with a liquid light guide to an inverted micro-

scope Leica DM-LFSA. All data collected from whole-cell recordings. Recordings were made using a MultiClamp700B amplifier

(Molecular Devices). Signals were filtered at 4 kHz and digitized at 10 kHz with a Digidata 1440A analog-digital interface (Molecular

Devices). pClamp10.6 software (Molecular Devices) was used to record and analyze data. Peak photocurrents weremeasured from a

1 or 5 s light pulse in voltage-clamp mode.

Cultured Neuron Calcium ImagingCalcium imaging conducted on neurons prepared and transfected as above in Tyrode media equivalent to primary culture electro-

physiology media, including blockers, except with (mM) 1 MgCl2 and 3 CaCl2. A custom-designed and 3D-printed stimulation insert

with platinum wires 1 cm apart was attached to a bipolar stimulator (Warner Instruments) time-locked to imaging software

(MetaMorph) via LabVIEW (National Instruments). Stimulation was conducted a 100 mA, 3ms pulse width, 0.1 Hz, with 5 pulses

per imaging field and images acquired at 33 Hz. Bulk signal was extracted from manually drawn ROIs and signals were corrected

for background fluorescence and GECI signal bleaching prior to calculating dF/F using custom MATLAB code. Baseline F was

defined as the median of the first 50 frames (approximately 1.5 s).


llNeuroResource


Virus productionAAV-8 (Y733F), was produced by the Stanford Neuroscience Gene Vector and Virus Core. In brief, AAV-8 was produced by standard

triple transfection of AAV 293 cells (Agilent). At 72 h post transfection, the cells were collected and lysed by a freeze-thaw procedure.

Viral particles were then purified by an iodixanol step-gradient ultracentrifugationmethod. The iodixanol was diluted and the AAVwas

concentrated using a 100-kDa molecular mass–cutoff ultrafiltration device. Genomic titer was determined by quantitative PCR. All

viruses were tested in cultured neurons for expected expression patterns prior to use in vivo. Virus titers were obtained for every

new batch of virus produced and virus expression was checked in cultured neurons prior to injection in vivo. Viral dilutions were

calculated based on original titers given at the time of virus production in order to match the viral titer of all viruses used in an exper-

iment. Aliquots of each virus were kept separately from in-use supply and, after all experiments were completed, all viruses used in

the Resource were re-titered simultaneously (across two plates) to standardize values across experiments and time and these final

re-titered values, and these values are reported in the submitted text, figures, and summary below. As virus titers vary based on

probe, standard curve, freeze-thaw cycles, and titer batch, we presented order-of-magnitude rounded titers in the figures and

text. Viral titers by figure rounded to nearest order of magnitude are below. Additional details are available in the Key Re-

sources Table.

Figures 3F, 3H, 3I, S6E, and S6F AAV8-Cre e10

AAV8-EYFP e12 AAV8-Flp e12

AAV8-Con/Fon-EYFP e12 AAV8-VCre e13

AAV8-Con/Foff-EYFP e11

AAV8-Coff/Fon-EYFP e12 Figure 5j

AAV8-Cre e10, e11, e12 AAV8-3x-G6M e12

AAV8-Flp e12 AAV8-Flp e10

AAV8-Flp-2a-Cre e12 AAV8-VCre e11

Figure 4B Figure S2B

AAV8-cDIO-EYFP e12 AAV8-jRGECO1a e12

AAV8-fDIO-EYFP e12 AAV8-sRGECO e12

AAV8-vcDIO-EYFP e12

AAV8-Cre e11 Figure S4F

AAV8-Flp e11 AAV8-Arch-EYFP e12

AAV8-VCre e12 AAV8-Con/Fon-Arch-EYFP e12

AAV8-Con/Foff-Arch-EYFP e12

Figure 4c AAV8-Coff/Fon-Arch-EYFP e12

AAV8-fDIO-EYFP e12 AAV8-Cre e12

AAV8-vcDIO-EYFP e12 AAV8-Flp e12

AAVretro-Flp e13 AAV8-Flp-2a-Cre e12

AAVretro-VCre e13

Figures S6A–S6D

Figure 5f AAV8-Con/Foff-EYFP e11

AAV8-3x-EYFP e12 AAV8-Flp e12

Stereotactic injectionsStereotactic viral injections were carried using typical technique. Briefly, mice induced andmaintained on isoflurane anesthesia were

placed in a stereotactic frame (Kopf Instruments) and the head leveled using bregma and lambda skull landmarks. Craniotomies were

performed so as to cause minimal damage to cortical tissue using a hand drill. Injections were made using a 10 mL syringe and 33 g-

36 g beveled needle (World Precision Instruments). 1000 nL of viral suspension was infused at a rate of 100 nl/min at the indicated

locations (coordinates in mm below). The needle was left in place for 10 minutes after the completion of injection before being with-

drawn under supervision. Skin was approximated with suture.


llNeuroResource


Stereotactic coordinates (mm) relative to Bregma

Anterior-Posterior Medial-Lateral Dorsal-Ventral

mPFC + 1.5 +/� 0.3 - 2.5

mPFC + 1.85 +/� 0.3 - 2.5

VTA - 3.3 +/� 0.4 - 4.2

Slice electrophysiologyTo prepare coronal slices (300 mm) frommice four weeks after viral injection, subjects were first trans-cardially perfusedwith ice-cold,

NMDG-HEPES recovery solution (‘NMDG solution’; Ting et al., 2014; in mM): 93 NMDG, 2.5 KCl, 1.2 NaH2PO4, 30 NaHCO3, 20

HEPES, 25 glucose, 5 sodium ascorbate, 2 thiourea, 3 sodium pyruvate, 10 MgSO4, 0.4 CaCl2, adjusted to pH 7.3 – 7.4 with HCl/

NaOH. A vibratome with ice-cold, bubbled (5% CO2 ‘carbogen’) NMDG solution was used to cut slices, which were then recovered

for 12 - 14 minutes in carbogen-bubbled NMDG solution at 32�C before being recovered for 1 hr in RT (22- 25�C) carbogen-bubbledartificial cerebrospinal fluid (‘aCSF; in mM): 124 NaCl, 2.5 KCl, 1.2 NaH2PO4, 24 NaHCO3, 5 HEPES, 12.5 glucose, 2MgSO4, 2 CaCl2,

adjusted to pH 7.3-7.4 with NaOH. Synaptic transmission blockers APV (25 mM), CNQX (10 mM) and sodium channel blocker TTX

(1 mM) are used as indicated. Electrophysiological recordings were performed at RT. Slices were visualized with an upright micro-

scope LEICA DM LFSA equipped with a 40x water-immersion objective. Individual neuron recordings were obtained after identifying

fluorescent protein expression without interruption of ACSF perfusion. Filtered light from a Spectra X Light engine (Lumencor) was

coupled to the fluorescence port of the microscope and used to both view fluorescence and deliver light pulses for opsin activation.

Whole-cell recordings were obtained with patch pipettes pulled from borosilicate glass capillaries (Sutter Instruments) with a hori-

zontal puller (P-2000, Sutter Instruments) and contained the following internal solution (in mM): 130 K-gluconate, 10 KCl, 10 HEPES,

10 EGTA, 2MgCl2, to pH 7.25with KOH. Recordings weremade using aMultiClamp700B amplifier (Molecular Devices). Signals were

filtered at 4 kHz and digitized at 10 kHz with a Digidata 1440A analog-digital interface (Molecular Devices). pClamp10.6 software

(Molecular Devices) was used to record and analyze data. Peak photocurrents were measured from a 1 s light pulse in voltage-clamp

mode.

Spectrophotometry and Viral Kinetic AnalysisA complete parts list is available in the Key Resources Table. Briefly, data describing viral expression kinetics in vivo were collected

with an inexpensive device consisting of a 505 nmLED (ThorlabsM505F1) coupled to a dichroic filtermount viamulti-mode fiber optic

patch cable (Thorlabs m76L01). Excitation light was band-pass filtered (497 nm/16 nm FWHM, Thorlabs MF497-16) and reflected by

dichroic (525 nm long-pass, Chroma T525lpxr) into a 200 mm, 0.53 NA (Doric) fiber optic patch cable to the sample. Emission light

then passed through the same dichroic, then through two identical clean-up filters (535 nm/22 nm FWHM, Thorlabs MF535-22), and

to a visible wavelength, compact CCD spectrometer (Thorlabs CCS100) via a round-to-linear fiber bundle (Thorlabs BFL200LS02).

The spectrometer was connected by USB to a computer running Windows and data acquired with the bundled Thorlabs spectrom-

eter software.

EYFP expression data were acquired from individual animals injected with various AAV8 viruses as indicated (Figure 3) and fitted

with 200 mm, 0.53 NA fiber optic implants (Doric) with the fiber tip placed at the injection site. Animals were injected on Wednesdays

(day 0) and the first reading taken on Friday (day 2) with (in ms), 5, 10, 50, 100, 500, 1000, 3000, 5000, 10000, and 20000 integration

times, with a large range designed to be sensitive to low expression (longer integration time), but also not saturated with high expres-

sion (shorter integration time). Readings were taken every Monday, Wednesday, and Friday thereafter for 6 weeks (18 time points).

Animals were sacrificed after time point 18 for expression analysis by confocal. Prior to recording each day, the excitation LED was

calibrated to 0.214mWusing a power meter (Thorlabs) and readings from a fluorescein slide (Thorlabs) and purified EYFPwere taken

to ensure consistent system performance. The raw data were stored as .txt files.

To calculate the ‘expression score’ of an individual time point, raw data were imported into MATLAB, signal from 528 nm - 540 nm

were segmented from the dataset, and the area under the curve was calculated using the trapz function across this signal subset for

each integration time point. Time points with zero or saturated values were excluded. Remaining values were normalized to integra-

tion time and averaged, resulting in that day’s expression score.

Virus-dependent expression kinetics were modeled in MATLAB by first log-transforming individual animal expression score data-

sets, normalizing these log-transformed data to the subject maximum, and pooling the samples within a condition. This combined

dataset was then fit with the equation y = 1*(1-exp(-b*x)) where y is fraction of max expression, x is the days of expression, and b is

solved for using the MATLAB fit function. Days of expression to a certain fraction of normalized log maximum expression was calcu-

lated from fitted curves by solving for x (days of expression) for a given y (fraction of maximum). 95% confidence interval of the fit and

the solved days of expression were calculated by using the 95% CI values of b.


llNeuroResource


To calculate absolute photon number, we calculated a conversion factor between photons and AUC (calculated as above) of a

fluorescein slide. To do this, we set the 505 nm excitation LED to 0.214 mW (mJ/s) at the tip of a fiber and connected this fiber to

the slide (as above). We next acquired spectrometry data for a series of integration time points between 100 ms and 500 ms, which

gave non-zero and non-saturated results. We then disconnected the patch cable (see Key Resources Table) connecting the optical

components to the spectrometer and measured light power coming out of this cable (and thus into the spectrometer) in the absence

of ambient light (with the power meter reading ‘zero’) using a highly sensitive power meter (Thorlabs CS130C) tuned to 534 nm (the

median wavelength within our filtered emission range), which was 2.0 nW. We calculated the photon energy (in J/photon) using the

equation E = hc/l, where h is Planck’s constant, c is the speed of light, and l is the photon wavelength, which we set at 534 nm. We

next solved for photons/second using E and the measured output emission power in W. Last, we scaled the acquired AUC values to

AUC/second and used this set of values to calculate a scale factor x for the spectrometer using the equation photons/s = x*AUC/s. To

extrapolate this scale factor to absorbance from AUC, we used absorbance values for a single wavelength (534 nm).

Fiber PhotometryMice were injected in VTA with combinations of 3x-GCaMP6m and recombinase AAV and fitted with 400 mm, 0.66 NA fiber optic

implants (Doric). After four weeks of expression, bulk fluorescence was collected as previously described (Gunaydin et al., 2014).

Calcium signal was collected by recording bulk fluorescence using a single optical fiber while simultaneously delivering excitation

light as previously described (Gunaydin et al., 2014; Inoue et al., 2019). We used a 385 nm LED (M385F1, Thorlabs) for movement

correction and 490 nm LED (M490F3, Thorlabs) for calcium signal recording. LEDS were filtered with 386-23 nm (FF01-386/23-

25, Semrock) and 488-10 nm (FF01-488/10-25, Semrock) bandpass filters. LED beams were combined using a 425 nm longpass

dichroic mirror (T425lpxr, Chroma) before being coupled into an optical fiber patch cable (400 nm diameter, 0.66 NA, Doric Lenses)

using a multiband dichroic (ZT405/488/561rpc, Chroma). The far end of the patch cord is end-end coupled to the fiber implant in the

animal using 2.5 mm ferrules and zirconia or bronze sleeves. Fluorescent calcium signal emission light was passed through a 555 nm

dichroic mirror (FF555-Di03-25x36, Semrock) which was filtered through a 447/522 nm dual band filter (ZET405/488 nm-custom nar-

row green, Chroma) and then focused onto a femtowatt photoreceiver (Model 2151, Newport) using a lens (62-561, Edmund Optics);

note that this system is designed to additionally collect red light and that this feature was not used for these experiments. The signal

was sampled at 6.1 kHz and independent signals were recovered using synchronous demodulation techniques, low-pass filtered

(corner frequency of 15 Hz, decimated to 382 Hz, then recorded to disk. Calcium signal was calculated for each continuous behav-

ioral recording with custom written MATLAB scripts. First, the 405 nm signal was subtracted from the calcium signal for motion

correction. Next, a double exponential was fitted to a thresholded version of the fluorescence time series and the best fit was sub-

tracted from the un-thresholded signal to account for slow bleaching. The fluorescence signal was normalized within each mouse by

calculating the dF/F as (F median (F)) / median (F), where the median was taken over the first 100 s of the trial. To shuffle bouts, the

bout structure (length of each bout and time between bouts) was preserved, while the starting time of the bouts was randomly placed

within the period of the trials when novel objects were present in the arena. When this caused bouts to extend beyond the end of the

trial, those bouts were wrapped to the beginning of the object interaction period, again preserving the bout structure. Each trial was

shuffled 300 times.

Novel object trials were conducted by placing subjects in a new cage after connecting them to the fiber photometry apparatus by

fiber optic patch cable (Doric). Behavior was recorded using an overhead mounted camera and synchronized with fiber photometry

using a trial-triggered LED that was mounted below the cage to be out of view of the subject. Animals explored the environment for

two minutes before novel objects were introduced. Animals were allowed to continue exploring for five additional minutes before the

trial was ended. Videos were scored manually for physical object interactions.

HistologyMice were trans-cardially perfused with 10 mL of ice-cold 1x PBS followed by 10 mL of 4% paraformaldehyde (PFA). After an over-

night post-fix in 4%PFA, brains were equilibrated in sterile 30% sucrose/1x PBS for at least 24 h (or until they sunk in the tube). Tissue

was sectioned at 60 mm using a freezing microtome (Leica) and mounted with DAPI-containing hard-mount solution (H-1500; Vector

Laboratories). Images were obtained on a Leica confocal microscope using 5x, 10x, 40x, and 63x objectives. For comparative

expression analysis, z stacks with the same settings (z-distance, number of optical slices, acquisition parameters) were taken

from the slice judged to have maximum expression. These were analyzed in Image-J (Fiji) by calculating the sum of the total inte-

grated fluorescence extracted from every optical section using a standardized ROI.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analysis was performedwith GraphPad Prism 7 andMATLAB. Statistical tests, number of subjects, and significance are as

detailed in the figure legends and text. General statistical approaches to analyzing data are as follows:

For two-sample comparisons of a single variable with sample sizes less than 10, we performed an independent, two-sample t test.

For simultaneous, pairwise comparisons of multiple, independent variables we performed one-way ANOVA with Sidak’s multiple

comparison test. For two-sample comparisons with individual sample sizes greater than 10, samples were first analyzed for distri-

bution normality using the Shapiro-Wilk normality test. If any samples from an experiment failed normality testing, comparisons of


llNeuroResource


unmatched data were performed with a Mann-Whitney test, while matched data comparisons were performed with a Wilcoxon

matched-pairs signed rank test; otherwise, an independent, two-sample t test was used.

Formultiway comparisons of a single variable with sample sizes less than 10, we performed a one-way ANOVAwith follow-up com-

parison of INTRSECT variants to parental (WT) sample using Dunnett’s multiple comparison test. For multiway comparisons with in-

dividual sample sizes greater than 10, samples were first analyzed for distribution using the Shapiro-Wilk normality test. If any sam-

ples from an experiment failed normality testing, comparisons were performed using a Kruskal-Wallis test with follow-up comparison

of INTRSECT variants to parental (WT) sample using Dunn’s multiple comparison test; otherwise, a one-way ANOVA was performed.

For comparisons across multiple variables with matched data, we performed two-way ANOVA with follow-up comparison of var-

iants to parental (WT) sample using Dunnett’s multiple comparison test.

To examine the linear relationship between related datasets, we first log-transformed data that spanned multiple log scales, then

analyzed matched datasets with normal distributions by computing the R2 value using Pearson correlation. Separately, to test the

dependency of a property on an experimental condition, we performed linear regression and examine the R2 value.

Viral expression kinetic curve fits are described in Spectrophotometry and Viral Kinetic Analysis.

Additional ResourcesDeisseroth Lab Website (listing of constructs, sequence maps, and viruses):


AddGene (repository for most constructs generated in this Resource):

http://www.addgene.org/ and see Key Resources Table

Stanford Viral Vector Core (Produced viruses in the Resource and maintains stocks of some of these viruses):


Flp mouse line database (living document version of Table 1):


INTRSECT Standard Operating Procedure (guidelines for experimental design):


NetGene2 (Splice site prediction software used for INTRSECT construct design):




http://www.addgene.org/





Documents

Comprehensive Dual- and Triple-Feature Intersectional Single ...web.stanford.edu/group/dlab/media/papers/fennoNeuron2020.pdfComprehensive Dual- and Triple-Feature Intersectional Single-Vector